A security researcher has added another dimension to smart lightbulbs by using them to store a library of banned books, creating what they describe as a “cyberpunk digital dead drop.” Rick Osgood wrote about the idea behind the Banned Book Library on his personal blog, alongside a retelling of the process to create a working prototype. Meanwhile, all the code is open source and available via Codeberg. So, anyone interested can acquire some ESP32-powered smart lightbulbs and distribute stealthy banned book libraries of their own.

Osgood notes that he was pondering making smart lightbulbs more useful. As someone “interested in infosec, open-source software, making things, breaking things,” he found inspiration in Ben Brown’s short story of anti-corporate tech resistance, called ‘Library.’ After that lightbulb moment, Osgood knew he wanted to make a small, inexpensive, cyberpunk‑style digital dead drop, designed for resilience and stealth – by reprogramming smart lightbulbs.

The security researcher chose a Tasmota‑flashed ESP32C3 smart bulb as the starting point for the stealth library project. As this device only has 4MB on board in total for everything, including the OS, and not much spare capacity left over, Osgood spent a considerable amount of time and effort wrangling with adding a microSD reader. However, this expansion effort proved fruitless, and thus he decided to live with the total capacity 4MB limit for the sake of sanity, convenience, and stealth.

Once Osgood accepted the 4MB storage limitation, a key task was to expand the existing ~320KB of free space for the filesystem as much as possible. After some optimizations and tuning, he managed to reserve about half of the total storage space available for banned books. We also assume some kind of text compression is used to make the best use of the limited capacity.

Thus, the current working prototype runs some custom ESP32 firmware, broadcasts an open Wi-Fi network, and hosts a server, with about half of the total capacity of the smart lightbulb left over. That means the all-important LittleFS library partition can be stuffed with almost 2MB of banned books. Osgood's example files use a selection of old public domain titles, presented in a CSS contents page. However, those who set up the lightbulbs will of course want to add their own digital tomes. Guidance for this, and for updating admin settings, and more is provided on the linked blog and Codeberg pages.

The Banned Book Library supports OTA updates. It is also worth mentioning that no cloud services are required, and no sensitive credentials need to be saved on the device. Please remember that modifying your smart lightbulb firmware could ‘brick’ it, so follow the linked guides at your own risk. Finally, also be careful carrying your banned book-stuffed smart lightbulbs when traveling internationally.

Google’s sunsetted Nest Gen 1 and Gen 2 thermostats have been given a new breath of life by a frustrated developer. The recently released No Longer Evil Thermostat project promises better control, with a “sleek, intuitive interface that rivals the original.” Moreover, it offers the benefit of community-driven software that is free and open source.

Nest Gen 1 and Gen 2 sunsetting backstory

These smart home thermostats were introduced between 2011 and 2014. However, they reached their end of support date on October 25 this year.

The impact of hitting the end of the support date is rather drastic for smart home proponents. With the passing of that date, the devices don’t just get shut off from firmware/software updates. They were also unpaired from the Nest and Google Home apps. That means the granular remote control and scheduling functionality evaporates. They still work as ‘dumb’ thermostats, though, on-site, with a click-wheel interface for heating adjustments.

No Longer Evil Thermostat to the rescue

Arizona-based security researcher and developer Cody Kociemba, part of the Hack House collaborative group, saw red when he became aware of Google’s plans to abandon the Nest Gen 1 and Gen 2 thermostats. Oiling the wheels of his hacking efforts somewhat was also the chance of a ~$15,000 bounty payment from the FULU Foundation. This organization basically crowdfunds financial rewards for devs who help liberate devices from corporate restrictions. Thus, he began to work on the open-source No Longer Evil Thermostat.

Kociemba says he is “passionate about hardware hacking, reverse engineering, and fighting corporate control,” so this project sounds like a great fit. A GitHub repository, which went live this week, provides all the tools and firmware those with abandoned Nest thermostats will need.

That link includes step-by-step instructions for getting the No Longer Evil Thermostat system up and running. If/when the process completes successfully, you will see the Nest boot and the “now made with 100% less evil!” welcome screen.

From this point on, you can use the No Longer Evil-infused thermostat with a local web interface, which uses an original Nest-a-like UI for temperature control, scheduling, and settings, etc. You can also integrate it with Home Assistant, if you wish, using MQTT or REST APIs. The option is also there to write a custom mobile wrapper for the UI.

Experimental / Testing warning

Before we go, we’d like to underline the No Longer Evil project developer's warning that, while this work offers the potential to “breathe new life into bricked and outdated Nest Generation 1 & 2 thermostats” and more, it is still in testing, and flashing this experimental firmware could brick your device. Please test this project only if you have a backup thermostat or heating system and can cope with a non-functioning device if things go wrong.

Cybersecurity researcher Dylan Ayrey of Truffle Security has shared a detailed blog post highlighting his experience with Eight Sleep smart beds since his discovery of an exposed AWS key inside of its firmware, prompting him to deeply investigate its security issues and find ways to alleviate them. Besides the AWS key problem, he also discovered a backdoor allowing SSH (Secure Shell) backdoor access and full arbitrary code execution capabilities, making Eight Sleep beds a disastrously unsafe device to keep on a home network for not just bed surveillance concerns, but the security of all devices involved.

Back in December, Ayrey made a Tweet from his @InsecureNature account encouraging his followers to guess what appliance of his had the major AWS key security issue, and this was before he even started talking about the SSH backdoor allowing arbitrary code execution on the bed.

Fast forward to now, and Dylan Ayrey has released an extended blog with the help of Jake King highlighting the security flaws of the Eight Sleep and the steps he ended up taking to make them no longer an issue, particularly in the face of features that wounded up locked behind a subscription paywall and Internet access for a bed that had already cost $2,000 to start.

According to Dylan, he was perfectly happy to deal with most of these downsides but still wound up curious about what might be hiding inside the firmware of Eight Sleep's temperature-controlled smart bed. His discovery gave him a serious case of "cyber ick" and prompted him to substitute the Eight Sleep pod otherwise used to regulate temperature with a regular aquarium chiller instead, which seemingly heats and cools the bed in the exact same way while only costing about $150. This involved cutting one of the tubes routed to the Eight Sleep pod and connecting it to an aquarium cooler instead, but proved a remarkably simple solution, providing "all the temperature control of an Eight Sleep with none of the apps, subscriptions, Internet connectivity, backdoors, and security liabilities of an Eight Sleep".

But what exactly are those security liabilities? Besides the exposed AWS key, which is mainly bad for reasons related to account security (though likely not the user's own, in this case), the biggest issue is backdoor SSH (Secure Shell) access. It seems that any of Eight Sleep's engineers can use SSH to access a customer's bed, detect when it's in or out of use, and execute whatever arbitrary code they please. While this mostly just means bed control and bed monitoring functionality when you limit your view to the bed itself, it gets much spookier when you consider that the smart bed is connected to the rest of your home network and thus jeopardizes those devices, too.

A serial burglar in Edina, Minnesota is suspected of using a Wi-Fi jammer to knock out connected security cameras before stealing and making off with the victim's prized possessions. Minnesota doesn’t generally have a reputation as a hotbed for technology, so readers shouldn’t be surprised to hear that reports of Wi-Fi jammers used to assist burglaries in the U.S. go back several years. PSA: even criminals use technology, and more are now catching on -- so homeowners should think about mitigations.

Edina police suspect that nine burglaries in the last six months have been undertaken with Wi-Fi jammer(s) deployed to ensure incriminating video evidence wasn’t available to investigators. The modus operandi of the thief or thieves is thought to be something like this:

• Homes in affluent areas are found
• Burglars carefully watch the homes
• The burglars avoid confrontation, so appear to wait until homes are empty
• Seizing the opportunity of an empty home, the burglars will deploy Wi-Fi jammer(s)
• “Safes, jewelry, and other high-end designer items,” are usually taken

A security expert interviewed by the source publication, KARE11, explained that the jammers simply confused wireless devices rather than blocking signals. They usually work by overloading wireless traffic “so that real traffic cannot get through,” the news site was told.

Jamming wireless security devices is a growing trend

Searching back through news reports concerning burglaries where Wi-Fi jammers have / may have been used surfaces plenty of prior evidence of this practice. A Ring community post about one of the firm’s wireless doorbells missing a porch thief after being subjected to a Wi-Fi deauthentication attack was posted back in January 2020. It is also easy to find reports of burglars using Wi-Fi jamming technology over 2021, 2022, and 2023 – with reports becoming more frequent over time.  

We mention Ring as one of the firms that popularized video doorbells to solve multiple home security concerns, but other wireless smart home security products from Blink (Amazon) and Nest (Google) will also be vulnerable to wireless signal jamming.

Worryingly, Wi-Fi jamming is almost a trivial activity for potential thieves in 2024. KARE11 notes that it could buy jammers online very easily and cheaply, with prices ranging from $40 to $1,000. Jammers are not legal to use in the U.S. but they are very easy to buy online.

Before we go, there are a few suggestions given to those now wondering about the efficacy of their home security systems with wireless components. Firstly, physically connect some of the devices which allow for a wired connection and local storage of footage. Secondly, utilize smart home technology that makes it appear that someone is at home. Your device may also have the ability to send alerts when the signal / connection is interrupted, and playing with those settings might be worthwhile.

An LG washing machine owner and self-confessed fintech geek has asked the Twitterverse why his smart home appliance ate an average of 3.66GB of data daily. Concerned about the washer’s internet addiction, Johnie forced the device to go cold turkey and blocked it using his router UI. Had the LG washer been hacked, hijacked, or otherwise tampered with over the net – or is this the average data consumption for a modern smart appliance?

Click 'See More' to expand the above Tweet / X.

Social media fans were eager to contribute to the detective work, sniffing out why the LG washing machine might be data-thirsty. Johnie’s initial screenshot showed that on a chosen day, the device uploaded 3.57GB and downloaded about 100MB, and the data traffic was almost constant. Meanwhile, according to the Asus router interface screenshot, the washing machine accounted for just shy of 5% of Johnie’s internet traffic daily.

Was the device downloading DLCs or the victim of a hack?

The LG washing machine owner saw the fun in his predicament and joked that the device might use Wi-Fi for “DLCs (Downloadable Laundry Cycles).” He wasn’t entirely kidding: The machine does download presets for various types of apparel. However, the lion’s share of the data transferred was uploaded.

Working through the thread, we note that Johnie also pondered the possibility of someone using his washing machine for crypto mining. “I’d gladly rent our LPU (Laundry Processing Unit) by the hour,” he quipped. Again, there was the glimmer of a possibility that there could be truth behind this joke. Another social media user highlighted a history of hackers taking over LG smart-connected appliances. The SmartThinQ home appliances HomeHack vulnerability was patched several weeks after being made public. A similar modern hack might use the washing machine’s computer resources as part of a botnet. Taking control of an LG washing machine as part of a large botnet for cryptocurrency mining or nefarious networking purposes wouldn’t be as far-fetched as it sounds. Large numbers of relatively low-power devices can be formidable together.

One of the more innocent theories regarding the significant data uploads suggested laundry data was being uploaded to LG so it could improve its LLM (Large Laundry Model). It sought to do this to prepare for the launch of its latest “AI washer-dryer combo” at CES, joked Johnie.

An Asus router issue?

For now, it looks like the favored answer to the data mystery is to blame Asus for misreporting it. We may never know what happened with Johnie, who is now running his LG washing machine offline. LG did not immediately reply to a request by Tom's Hardware for more information.

Another relatively innocent reason for the supposed high volume of uploads could be an error in the Asus router firmware. In a follow-up post a day after his initial Tweet, Johnie noted “inaccuracy in the ASUS router tool,” with regard to Apple iMessage data use.  Other LG smart washing machine users showed device data use from their router UIs. It turns out that these appliances more typically use less than 1MB per day.

The real dangers of connected device hacking

Though Johnie’s Tweets were relatively light-hearted, hacking smart connected devices can be severe. Consider what could happen if medical or industrial IoT devices get taken over by attackers, for example. A case in point is provided by a story from earlier this week when Bosch network-connected wrenches used in factories all around the world were found to be riddled with vulnerabilities.

Researchers highlighted that the cordless industrial wrenches could be hit by exploits or ransomware – with the threat of turning off the wrenches en masse. The wrenches could even be secretly readjusted to make things they were used to construct a danger to their users – through applying incorrect torque. Patches are due soon.

Samsung looks set to bet big on the appeal of AI enhancements in its wide range of home appliances. According to a report published by DigiTimes earlier today, the South Korean tech giant has plans to "equip all its new home appliance products with neural processing units (NPUs)," in 2024. The plans encompass both premium and mainstream appliances in product categories like Smart TVs, dishwashers, ovens, refrigerators, and air conditioners. Who knows, it might even add AI to toasters.

The reason for Samsung's investment in equipping NPUs in such a widespread manner is to add artificial intelligence (AI) to its appliances as a key differentiator, indicates the source report. DigiTimes says its sources indicate the Samsung Home Appliances Division is busy working on updated chipsets. Ultimately, the goal is to enable appliances with ultra-low power use, but always-on AI smarts, to assist customers.

Some of the specific tasks the AI built-into your home appliance may be to leverage include: on-device AI processing to implement voice recognition, machine vision, and full enabling of the Bixby assistant. Bixby will add question-and-answer capabilities to the devices, and more widely integrate all your compatible home devices.

Some recently showcased smart home appliances from the likes of Samsung take things further. For example, ovens are becoming capable of recognizing food being cooked and will suggest cooking methods. Another example mentioned in the DigiTimes report is of refrigerators recognizing stored items.

Samsung's 2024 plans were also illuminated by information from NPU providers like Korea's DeepX. The DigiTimes report says that DeepX is planning to mass produce its next-gen NPU in Q3 next year, with the goal of eradicating Chinese-made competitor components.

If we look more closely at Samsung's recent press releases from IFA 2023, like this one published this morning, we can see more clearly why adding NPUs to everything is Samsung's latest product trend. More smart appliances from Samsung will help drive adoption of its SmartThings ecosystem, which will help create a viscous circle for the company, driving adoptiooin of its own (and partner) products.

Samsung's examples of SmartThings Routines, where home appliances work together and with the home owner for greater convenience and efficiency, sound quite appealing. In one example, it suggests creating a 'Sunshine Mode' in SmartThings, where the air conditioner is activated on sunny and hot days, when solar panels are getting plenty of charge. Moreover, SmartThings compatible curtains and blinds can help make the indoor environment more comfortable.

Adding AI to home appliances won't be welcomed by everyone, though. Some will see security and privacy concerns in home appliances with voice recognition, machine vision, and the ability to 'think' for themselves.

When it comes to creating a project for school, we’re super biased when we say that a Raspberry Pi project is probably your best bet. Engineering students Alexander Calvert and Nathan Ferguson of Monash Engineering have created an impressive quadruped robot called the Dingo that relies on our favorite SBC, the Raspberry Pi. According to the students, the intention was to create a low-cost solution that would be ideal for research and modifiable with additional components.

The robot is capable of locomotion with varying degrees of control. You can operate its pitch, roll, and yaw and adjust for speed. It can crouch and even carry a little bit of weight to transport things. It’s remote-controllable, as well. In this case, the duo is using a Play Station 4 controller. The team also explains that this project is a fork of Stanford’s Quadruped project which can be found over at GitHub.

Although the team was aiming for a low-cost robot, it still comes with a hefty price that might put off some hobbyists. They provided a full list of parts that totals over $1,300 USD. However, you can replace some of the components with cheaper alternatives but the performance quality will likely drop in some areas. The body is 3D-printable and totally open source for anyone who wants to print their own or modify its design. You can find the files over at Grab CAD.

The main board used to control the Dingo is a Raspberry Pi 4 B. It’s assisted by an Arduino Nano and a huge list of hardware which includes 12 servos priced at $44.49 a piece (that adds up to over $530 USD worth of servos alone). You can peruse the complete parts list on the project’s GitHub page.

The main operating system used in the build is Ubuntu. You’ll need a few extra tools to bring it together including VSCode and ROS Noetic. There are also some custom Python scripts to handle things like controller input. Check out the official Dingo Quadruped GitHub to get a closer look at the software side of its design.

If you want to see this Raspberry Pi project in action, check out the demo video uploaded by the students to YouTube and give them some support for their hard work this semester. You might want to keep an eye out for these two for more future creations.

If one thing is for sure, you can’t have too many Raspberry Pis. And if you do have too many Pis for some reason, there’s no excuse for not cramming them all into one cool project. Today we’ve got an awesome creation by maker Alexander Rau who seems to have done just that. Rau is using eight Raspberry Pi CM4 modules to power his custom Kubernetes cluster.

Rau documented the process in a very thorough blog post, detailing how he went from the concept phase all the way up to the finished product. He knew he wanted to use K3s, a Kubernetes distribution, and he knew he wanted to create a custom PCB for the project but bringing his idea to fruition proved to be a challenging yet rewarding endeavor.

There were a few major hurdles in developing the project. First, developing a custom carrier PCB for the CM4 modules was a big undertaking. It needed to provide a few key features and existing carrier boards had more components than the project called for. He also needed to create a cluster tray from scratch to house his PCBs. On the software side, he needed to learn how to manage the hardware with Rancher and set up the system to be automated using Ansible.

After many months of planning and prototyping, he ended up with what he calls the Less-is-More carrier board, abbreviated to just LiM. This PCB supports the Raspberry Pi CM4 and has Ethernet support, a microSD slot, an M.2 M-key socket for SSDs, and more. The cluster tray was 3D printed and supports a total of 8 cluster nodes.

As we mentioned above, Rau opted to use Rancher to manage the cluster. Each of the Raspberry Pi CM4 modules are running the latest version of Raspberry Pi OS. This is necessary to automate the cluster using Ansible. Rau was kind enough to share plenty of details about the software set up in the original project post.

If you want to recreate this Raspberry Pi project or just get a closer look at how it goes together, check out the original post shared on Alexander Rau’s website. You can also follow him for more cool creations as well as any future updates on this neat cluster project.

According to a Tweet by Y-Cruncher author Alexander Yee, the soon to be release version 0.7.10 of the Y-Cruncher app will arrive with full AVX-512 support for AMD's upcoming Ryzen 7000 processors. But apparently the work to add Zen 4 support wasn't easy, Yee states he had to do all the optimizations by himself, with a 7950X sample he owns, since AMD has yet to share any optimization resources for Zen 4.

If you are unfamiliar with Y-Cruncher, its a PI calculation benchmark for multi-core processors that has been around for quite a while. Similar to Prime95, Y-Cruncher calculates how fast your PC can generate Pi or other constants to many trillions of digits. This ability makes the application great for benchmarking and stress-testing computers.

We are not sure how Yee got his hands on a engineering sample of a Ryzen 9 7950X. But, if it was given to him by AMD ahead of launch, that is good news. If more developers are receiving or requesting engineering samples of Ryzen 7000 chips right now, it means Ryzen 7000 support for many apps and drivers will be ready in time for release. 

Ryzen 7000 is AMD's upcoming successor to the Ryzen 5000 series CPUs, running on AMD's bleeding edge Zen 4 architecture. IPC improvements are expected to be around 8% to 10% with the rest of Zen 4's performance improvements coming from much higher clock speeds compared to Zen 3, with an estimated 5.5 GHz average boost frequency for Ryzen 7000 chips.

Of course, one of the most surprising features being introduced with Zen 4 is the introduction of the previously Intel exclusive AVX-512 instruction set, which ironically, was officially removed from Intel's 12th Gen Alder Lake CPUs. Despite Intel being the original creator of AVX-512 and pushing hard for its adoption in the early days of its life.

Like vanilla AVX and AVX2, AVX-512 is another instruction set that aims to accelerate specific workloads with specialized silicon within the processor. To this day, its still a rarely used instruction set, with some people calling for its death. But for the few apps that do support it, it can make a big difference, a good example of this is PS3 emulation.

With AVX-512 coming to Zen 4, and developers like Alexander Yee focusing on supporting the new instruction set implementation on AMD, AVX-512 gets closer to becoming a mainstream item. Hopefully AMD's implementation is strong enough to incentivize Intel to reintroduce it back into its mainstream platforms officially.

Building the smart home of your dreams is as easy as Pi—Raspberry Pi that is. To prove this, today’s project created by Jithin over at the Electromaker website is a Raspberry Pi Pico powered project that controls a door lock using Amazon Alexa’s voice control system.

The project works just as you might expect. It listens for the user to make an audible request and parses the voice command using Amazon's Alexa servers which will trigger the door to lock or unlock based on what was said. This project requires internet access to work so Jithin used the first Wi-Fi enabled Pico, the $31 Arduino Nano RP2040 Connect board. That said, this project could easily be recreated using the new $6 Raspberry Pi Pico W at a much lower cost.

This isn’t Jithins first foray into microelectronics. According to their profile at Electromaker, Jithin has an affinity for making projects using the Raspberry Pi, Arduino boards, and coding with Python. Previous examples of their work include this soldering iron fume extractor with built-in lighting.

This project relies on a custom PCB designed by Jithin and printed by PCBWay. According to Jithin, the circuit was designed using an application called Altium Designer. A few extra components are needed, as well, including the Arduino Nano RP2040 Connect and a 4-channel relay.

The software for the voice control system was written using Arduino IoT Cloud and works in conjunction with the Alexa application available on smartphones. This app has a tool specifically designed to work with Arduino boards known as Arduino Skill which is necessary to communicate with the Arduino Nano RP2040 Connect on the PCB.

For a closer look at this Raspberry Pi project, visit the original project thread over at Electromaker for a complete breakdown of how to recreate the system at home. You can also find a demo video of the project over at YouTube. Be sure to follow Jithin for more creations as well as any future updates on this one.

The Raspberry Pi community has a rich history of integrating the classic SBC into a plethora of robotic projects and this new RP2040 motor driver board developed by maker Taylor Alexander is paving the way for even more awe-inspiring Pi-powered robots. The board was unveiled earlier this week over at Twitter with an up-close look at its open-source design.

This credit-card-sized board uses the RP2040 microcontroller as the main driver for up to two motors using encoders at 45 A per motor supporting up to 60 V. We reached out to Alexander who confirmed the RP2040 handles all of the motor control but gate drivers are used to protect against shoot-through which would short circuit the power supply.

Alexander has a history specializing in robotics with an affinity for DIY microelectronics-based projects. This RP2040 motor driver is just a part of that history and serves as a new addition to a previous robot we covered for automated farming known as Acorn. What’s cool about the new board is that it’s not only a new development for the Acorn project but also an open-source tool for anyone who wants to integrate it into their own project.

According to Alexander, a big goal with its design was to use components that are easy to source so makers could order it as one unit without having to obtain parts separately. Everything from the gate drivers to the MOSFETS are all obtainable through JLCPCB but the raw files are available for anyone who wants to order the board through any PCB manufacturer of their choosing.

The RP2040 motor controller was designed using KiCad, an open-source application designed for makers to create PCBs from scratch, but the file can be edited using other applications as well. Users can get a closer look at the board design and associated files on the official GitHub page. In addition to its integration with Acorn, Alexander also plans to use it as a driver for this impressive 3D-printed four-axis robot arm project.

If you want to use this board for your own Raspberry Pi project or just get a closer look at its design, check out the original thread shared to Twitter. This project, as well as Acorn, was funded by Daniel Theobald who has a blog known as Twisted Fields which highlights more farming-based robotics projects. Be sure to follow them both for more updates and cool developments in the future.

If you’re looking for an automated solution for your home bar, check out this Raspberry Pi-powered cocktail maker created by Alexander Liggesmeyer. It does exactly what it sounds like—it uses a Raspberry Pi to make cocktails at the push of a button.

The best Raspberry Pi projects are easy to share and this one features a notably user-friendly web-based interface. Users access the UI from a browser and choose the drink they want to order. The Pi then uses a series of pumps to create the drink algorithmically. So just be sure to have your glass ready!

According to Liggesmeyer, the project relies on a Raspberry Pi 4B, which connects to a relay board that helps monitor eight food-safe peristaltic dosing pumps and diodes. The drink ingredients are transported using sections of a silicone hose.

The UI is pre-programmed with a menu of drinks, but users can create their own recipes to share with others. Then, if all of the ingredients are available, the drink can be ordered.

To get a closer look at this Raspberry Pi-powered cocktail maker and see how it goes together, check out the official project page at GitHub and be sure to follow Alex Liggesmeyer for more cool creations.

If you’ve ever wanted to bring a little bit of 2001: A Space Odyssey into your house, then you should appreciate this Raspberry Pi-based Hal 9000 project by maker Cheat_Steve10. It looks just like Hal from the 1968 sci-fi classic, but, thankfully, it doesn’t quite function like the original.

Some of the best Raspberry Pi projects provide multiple features, and this Hal 9000 recreation does just that. The most notable function is its voice control support which operates with the help of Alexa Voice Service (AVS). In addition to talking back, it also acts as a Samba NAS device for managing data.

A Raspberry Pi 3B model was used as the mainboard for the project. A PAM8403 amplifier was necessary to get audio output, while a USB microphone allows audio to be recorded. An old lens was fitted into the housing with a red LED in the center to complete the look. According to Cheat_Steve10, he still had to make a cap for the lens because it was too eerie when left on at night.

Software-wise, the setup isn’t too complicated. Cheat_Steve10 claimed to use Raspbian, but there’s no reason this configuration wouldn’t work with the latest Raspberry Pi OS. The AVS setup is handled by the AVS Device SDK Sample App, while the NAS tools are managed by Samba.

To get a closer look at this creepy (yet awesome) Pi project, check out the original Reddit thread and be sure to follow Cheat_Steve10 for more updates. 

In this tutorial, we’ll install Alexa Smart Screen SDK on a Raspberry Pi 4 to essentially “make” our own version of the Echo Show. We’ll nickname this project, “PiShow.” This project is possible with the Amazon Alexa Smart Screen SDK, in addition to the Alexa Voice Service (AVS) SDK discussed in our previous article, How to Build an Alexa Speaker with Raspberry Pi. 

Caveats about PiShow 

• PiShow is not intended to be a replacement for the Echo Show. While many of the Echo Show capabilities are included with this version, significant capabilities including the ability to play videos natively were intentionally excluded from this build.
• This smart screen version of AlexaPi does not include the vocal wake word trigger “Alexa.” The user will press the ‘A’ key when speaking to Alexa.
• PiShow cannot launch with VNC active. VNC must be disabled on the Raspberry Pi.
• Music - The user can pause music on a Pi touchscreen. Music capabilities are the same as described in the AlexaPi article.
• Alexa Skills are available on PiShow, and any visual screens within an Alexa Skill are also visible on PiShow. Furthermore,  touchscreen functionality is enabled for Alexa Skills.
• If you are using a touchscreen, and you tap the screen while Alexa is speaking, she will stop speaking. This is the same behavior on the Echo Show. 

Why Build a PiShow? 

In our previous article, we addressed the cost differential of purchasing the AlexaPi project components vs. an Echo Dot. In this post, we fully disclose that it is less expensive to purchase an Echo Show vs. the components for PiShow. 

If you already own a Raspberry Pi and a touchscreen, this could be a fun weekend STEM project. The fun is in the making and learning about how voice technology works behind the scenes.

What You’ll Need

• Raspberry Pi 3B+ or Raspberry Pi 
• 16GB (or larger) microSD card (see best Raspberry Pi microSD cards) with a fresh install of Raspberry Pi OS
• Power supply/Keyboard/Mouse/Monitor/HDMI Cable (for your Raspberry Pi)
• USB Microphone
• Speaker with 3.5mm or USB connector. We’ve only tested with a 3.5mm speaker but USB should work also.
• HDMI monitor, 7” Raspberry Pi Touchscreen, or 3.5” or 5” TFT screen - This project works with touchscreen capabilities.
• Optional: Mini wireless keyboard - useful for starting the script and holding down the ‘A’ key after install.

Timing: Plan for a minimum of 3 hours to complete this project. The AVS Device SDK make install step takes around 2 hours depending on your Pi model and internet speed.

Prerequisites:

Before starting this tutorial, complete the entire project as detailed in our previous article, How to Build an Alexa Speaker with Raspberry Pi.

The majority of this tutorial is based on terminal commands. If you are not familiar with terminal commands on your Raspberry Pi, we highly recommend reviewing 25+ Linux Commands Raspberry Pi Users Need to Know first. 

PiShow Setup and Install 

1. Important: Complete the AVS installation as detailed in our previous article, How to Build an Alexa Speaker with Raspberry Pi.

2. If your AlexaPi is currently running, press Ctrl-C to stop the script. Speaker, mic and power should still be connected.

3. Attach your screen : I added a 7” Raspberry Pi Touchscreen for the PiShow version. But you could use a third-party touch screen that connects via either HDMI or via the GPIO pins. You can find such screens for as little as $23 on Amazon. If you use a screen that’s non-touch, you can’t use those features.

Back view:

4. Open a Terminal.

5. We will start by creating the folders and installing the libraries and apps we will need for PiShow. Enter the following commands.

cd /home/pimkdir sdk_foldercd sdk_foldermkdir sdk-build sdk-source third-party sdk-install dbsudo apt-get -y install \git gcc cmake build-essential libsqlite3-dev libcurl4-openssl-dev libfaad-dev \libssl-dev libsoup2.4-dev libgcrypt20-dev libgstreamer-plugins-bad1.0-dev \   gstreamer1.0-plugins-good libasound2-dev doxygenpip install commentjson

Your Pi may already have commentjson installed. 

6. Next, we will install PortAudio to record microphone data.

cd third-partywget -c http://www.portaudio.com/archives/pa_stable_v190600_20161030.tgztar zxf pa_stable_v190600_20161030.tgzcd portaudio./configure --without-jack

7. Run ‘make’ in the PortAudio folder. This command took around 1.5 minutes on a Raspberry Pi 4.

make

cd $HOME/sdk_folder/sdk-sourcegit clone --single-branch --branch v1.21.0 git://github.com/alexa/avs-device-sdk.gitcd $HOME/sdk_folder/sdk-buildcmake $HOME/sdk_folder/sdk-source/avs-device-sdk \-DGSTREAMER_MEDIA_PLAYER=ON \-DPORTAUDIO=ON \-DPORTAUDIO_LIB_PATH=$HOME/sdk_folder/third-party/portaudio/lib/.libs/libportaudio.a \-DPORTAUDIO_INCLUDE_DIR=$HOME/sdk_folder/third-party/portaudio/include \-DCMAKE_BUILD_TYPE=DEBUG \-DCMAKE_INSTALL_PREFIX=$HOME/sdk_folder/sdk-install \-DRAPIDJSON_MEM_OPTIMIZATION=OFFmake install

When the make install is complete, there should be no error messages on your screen.

9. Copy the config.json file from the AlexaPi project (/home/pi/) to the Install folder and generate the AlexaClientSDKConfig.json file.

cp ~/config.json $HOME/sdk_folder/sdk-source/avs-device-sdk/tools/Installcd $HOME/sdk_folder/sdk-source/avs-device-sdk/tools/Installbash genConfig.sh config.json \your-device-serial-number \$HOME/sdk_folder/db \$HOME/sdk_folder/sdk-source/avs-device-sdk \$HOME/sdk_folder/sdk-build/Integration/AlexaClientSDKConfig.json \-DSDK_CONFIG_MANUFACTURER_NAME="manufacturer name" \-DSDK_CONFIG_DEVICE_DESCRIPTION="device description"

10. Open File Manager and navigate to /home/pi/sdk_folder/sdk-build/Integration/ folder.  

11. Right-click on AlexaClientSDKConfig.json and select Text Editor to open the file for editing.

12. Add the info for gstreamerMediaPlayer in the first set of brackets.

"gstreamerMediaPlayer":{ "audioSink":"alsasink"},

13. Save your file. 

14. You can also save a backup copy as AlexaClientSDKConfig_backup.json 

15. Close your Text Editor.

16. Create or modify your /.asoundrc file.

cdsudo nano ~/.asoundrc

17.  Add the following lines to ~/.asoundrc 

pcm.!default {   type asym   playback.pcm {     type plug     slave.pcm "hw:0,0"   }   capture.pcm {     type plug      slave.pcm "hw:1,0"   } }

18. Press Ctrl-X, Y, and Enter to save your ~/.asoundrc file.

19. Test your progress so far.

cd $HOME/sdk_folder/sdk-build PA_ALSA_PLUGHW=1 ./SampleApp/src/SampleApp ./Integration/AlexaClientSDKConfig.json DEBUG9

20. Similar to the authentication you completed during the AlexaPi project, in the Terminal, scroll up to find your code, then navigate to http://amazon.com/us/code and enter your code.  

21. Press Continue and you should see a Success message in your browser. 

22. Go back to your Terminal and scroll up to search for the “Authorized” message. 

23. This AlexaPi works slightly differently from the original AlexaPi project in that you’ll have to type ‘t’ followed by the Enter key before speaking to Alexa. Try it now. Press ‘t’ enter and say, “What time is it?”

If Alexa provides the time, then you have successfully completed installing the AVS Device SDK sample app. Congratulations!

24. Press Ctrl-C to stop the sample app.

25. Next, we will download and ‘make’ the APL Core Library. The last command in this sequence ‘make’ took approximately 15 minutes to complete.  

cd $HOME/sdk_folder git clone --single-branch --branch v1.4.1 git://github.com/alexa/apl-core-library.gitcd $HOME/sdk_folder/apl-core-librarymkdir buildcd buildcmake ..make

26. Install Alexa Smart Screen SDK dependencies, Websocket++, AISO and Node.js

cd $HOME/sdk_folder/third-partywget https://github.com/zaphoyd/websocketpp/archive/0.8.1.tar.gz -O websocketpp-0.8.1.tar.gztar -xvzf websocketpp-0.8.1.tar.gzcd $HOME/sdk_folder/third-partysudo apt-get -y install libasio-dev --no-install-recommends  cd $HOME/sdk_folder/third-partycurl -sL https://deb.nodesource.com/setup_13.x | sudo -E bash -sudo apt-get install -y nodejs

27. Download and install Alexa Smart Screen SDK. The last step ‘make’ took approximately 25 minutes to complete.

cd $HOME/sdk_folder     git clone git://github.com/alexa/alexa-smart-screen-sdk.gitmkdir ss-buildcd ss-buildcmake -DCMAKE_PREFIX_PATH=$HOME/sdk_folder/sdk-install \ -DWEBSOCKETPP_INCLUDE_DIR=$HOME/sdk_folder/third-party/websocketpp-0.8.1 \ -DDISABLE_WEBSOCKET_SSL=ON \ -DGSTREAMER_MEDIA_PLAYER=ON \ -DCMAKE_BUILD_TYPE=DEBUG \ -DPORTAUDIO=ON -DPORTAUDIO_LIB_PATH=$HOME/sdk_folder/third-party/portaudio/lib/.libs/libportaudio.a \ -DPORTAUDIO_INCLUDE_DIR=$HOME/sdk_folder/third-party/portaudio/include/ \ -DAPL_CORE=ON \ -DAPLCORE_INCLUDE_DIR=$HOME/sdk_folder/apl-core-library/aplcore/include \ -DAPLCORE_LIB_DIR=$HOME/sdk_folder/apl-core-library/build/aplcore \-DAPLCORE_RAPIDJSON_INCLUDE_DIR=$HOME/sdk_folder/apl-core-library/build/rapidjson-prefix/src/rapidjson/include \ -DYOGA_INCLUDE_DIR=$HOME/sdk_folder/apl-core-library/build/yoga-prefix/src/yoga \ -DYOGA_LIB_DIR=$HOME/sdk_folder/apl-core-library/build/lib \  ../alexa-smart-screen-sdkmake

Next, we can test our PiShow app. (If you are using VNC, you will need to stop and disable VNC in order for the app to launch.) 

Run PiShow 

28. Open File Manager and navigate to: /home/pi/sdk_folder/ss-build/modules/GUI/index.html 

29. Open index.html with your Chromium browser by double-clicking on index.html 

30. Go back to your Terminal and enter the following commands to start your PiShow. 

cd $HOME/sdk_folder/ss-build PA_ALSA_PLUGHW=1 ./modules/Alexa/SampleApp/src/SampleApp -C \ $HOME/sdk_folder/sdk-build/Integration/AlexaClientSDKConfig.json -C \ $HOME/sdk_folder/alexa-smart-screen-sdk/modules/GUI/config/SmartScreenSDKConfig.json -L INFO

31. You may need to authenticate the PiShow sample app as you did during step 22. If so,  scroll up to find your code in the Terminal, then navigate to http://amazon.com/us/code and enter your code. 

32. Go back to your Chromium browser and you should see the message “Press and Hold “A” then Speak” 

33. Give it a try! Quick note: index.html in Chromium must be the active window for PiShow to work. Hold down the ‘A’ key while you say, “Tell me the weather.”

Congratulations! You have made your PiShow!

Always use the same process: With /home/pi/sdk_folder/ss-build/modules/GUI/index.html as the active window, hold down the ‘A’ key while speaking to Alexa. You do not need to say the wake word, “Alexa.” 

Things to Try on your new PiShow 

• “Play Music” - If using the touchscreen, you can pause the music by tapping the pause button on the screen.
• “Tell me a joke.”
• “Open Big Sky” - Weather app with great graphics.
• “What does Planet Earth look like?”
• “Enable Space Station”

To stop PiShow, go back to your Terminal and press Ctrl-C. 

If you turn off or reboot your Pi, you can restart PiShow with the Terminal command: 

cd $HOME/sdk_folder/ss-build PA_ALSA_PLUGHW=1 ./modules/Alexa/SampleApp/src/SampleApp -C \ $HOME/sdk_folder/sdk-build/Integration/AlexaClientSDKConfig.json -C \ $HOME/sdk_folder/alexa-smart-screen-sdk/modules/GUI/config/SmartScreenSDKConfig.json -L INFO

In this tutorial, we’ll install Alexa Voice Service (AVS) on a Raspberry Pi 4 to essentially “make” our own Alexa speaker. We’ll nickname this project, “AlexaPi.” This project is possible thanks to the Amazon Alexa Voice Service (AVS) SDK. If you have your heart set on an Echo Show on Raspberry Pi (Alexa with a screen), we’ll cover a Raspberry Pi Alexa with a visual interface in a future article. Completing the steps in the article is a prerequisite to the AlexaPi with screen project. 

Why Build an Alexa when you could just buy one? 

If you don’t already own a Raspberry Pi, it is less expensive to buy an Amazon Echo device. As of this writing, a brand new Echo Dot sells for $24.99 on Amazon’s website which is $10 cheaper than the cheapest Raspberry Pi 4 and that’s not including the cost of other components.

The joy of AlexaPi is not the outcome, but the educational journey. You’ll get a glimpse of how Alexa works behind the scenes, and possibly spark an interest in designing your own custom device with Alexa embedded. Amazon has made this Pi project possible for the purpose of encouraging device manufacturers to embed Alexa in their products.

What is the difference between AlexaPi and Echo devices? 

AlexaPi has many of the basic features of an Echo device, but is not intended to be a replacement for purchasing an Echo device. Below are a few of the key similarities and differences between Echo and AlexaPi. 

• Music services enabled by default include iHeartRadio, TuneIn, and Audible (for US users, at the time of the writing of this article). According to Amazon, additional Music Service Providers (MSPs) must be requested with “your Amazon point of contact.” Once again, this project was created for hardware manufacturers to embed Alexa in their commercial products.
• Smart Home devices such as TP-Link Kasa, Teckin, Avatar, and more can be controlled via your AlexaPi with the same process as an Echo device.
• Alexa Skills are available on AlexaPi. For example, say, “Alexa open cat food.” This is a skill I wrote that plays the sound of opening a can of cat food.
• Calling & Messaging - At the time of the writing of this article in the US, calling via AlexaPi is not supported.

What You’ll Need

• Raspberry Pi 3 or Raspberry Pi 4
• 16GB (or larger) microSD card (see best Raspberry Pi microSD cards) with a fresh install of Raspberry Pi OS
• Power supply/Keyboard/Mouse/Monitor/HDMI Cable (for your Raspberry Pi)
• USB Microphone
• Speaker with 3.5mm or USB connector. We’ve only tested with a 3.5mm speaker but USB should also work.

Timing: Plan for a minimum of 1 hour to complete this project. The main install step takes 30 minutes to over an hour depending on your Pi model and internet speed.

The majority of this tutorial is based on terminal commands. If you are not familiar with terminal commands on your Raspberry Pi, we highly recommend reviewing 25+ Linux Commands Raspberry Pi Users Need to Know first.  

AlexaPi Setup and Install 

1. Connect your USB microphone and speaker to your Raspberry Pi.

2. Boot your Raspberry Pi. If you don’t already have a microSD card see our article on how to set up a Raspberry Pi for the first time or how to do a headless Raspberry Pi install.

3. Launch Chromium browser on your Pi desktop.

4. Create or login to your Amazon developer account at: https://developer.amazon.com/

5. Click on Developer Console in the top right corner.

6. Mouse over Alexa and click Alexa Voice Service.

7. Click MANAGE YOUR PRODUCTS 

8. Click ADD NEW PRODUCT.

9. Complete the Product Information form.

• Give your product a name and ID.
• Select Device with Alexa built-in for product type.
• Select No for companion app.
• Select Other for Product Category and enter your value in the text field.
• Enter your Brief product description.
• Select Hands-free for “How will end users interact with your product?”
• Uploading an image is optional.
• Select No for commercial distribution, Alexa for Business, AWS IoT Core, and children’s product.
• Click Next.

Next, you’ll be prompted to set up your AVS Security profile.

10. Click CREATE NEW PROFILE

11. Enter your profile name and description.

12. Click Next

13. Click Other devices and Platforms in the Platform Information area

14. Enter a Client ID name of your choosing.

15. Click GENERATE ID

16. Click Download and a config.json file will download to your Pi. 

17. Check the box for terms and conditions and click Next. 

18. Click OK. You’ll be taken back to the product page. You should see the product you just created. 

19. On the top right corner of your Raspberry Pi desktop, right click on the speaker icon and select AV Jack. 

20. Left click on the same speaker icon and set the volume to 100%.

21. Open your file manager and navigate to your Downloads folder. 

 22. Move your config.json file from Downloads to pi. 

23. Open a Terminal on your Raspberry Pi desktop.

24. Run updates:

udo apt-get update && sudo apt-get upgrade

25. Next, enter the following as one command.

wget https://raw.githubusercontent.com/alexa/avs-device-sdk/master/tools/Install/setup.sh \wget https://raw.githubusercontent.com/alexa/avs-device-sdk/master/tools/Install/genConfig.sh \wget https://raw.githubusercontent.com/alexa/avs-device-sdk/master/tools/Install/pi.sh

26. Next, run the install script

sudo bash setup.sh config.json [-s 1234]

27. Type AGREE and hit Enter (if you agree with AVS Terms and conditions). This step can take 30 minutes to over an hour depending on your Raspberry Pi model and internet speed. 

28. During the install process you’ll be prompted to review the license agreement. Press Enter to view the agreement. Press the spacebar to page down through the agreement.

29. If you accept the license agreement, type yes when prompted. 

Your build is complete when you see **** Completed Configuration/Build ***

Troubleshooting

• In some cases, your Raspberry Pi may freeze and/or over heat during installation. If this happens to you, unplug your Raspberry Pi and allow it to cool for at least 5 minutes. Once your Pi has cooled, plug it back in and restart your terminal command sudo bash setup.sh config.json [-s 1234]
• If you get a fatal error: curl/curl.h: No such file or directory that means you need to install curl. In your Terminal enter:

sudo apt-get install libcurl4-openssl-devsudo apt-get install libcurl4-gnutls-dev

Launching AVS on Raspberry Pi 

If you reached **** Completed Configuration/Build *** in the previous step, Congratulations! In the next step, we will start our Alexa app, and Authenticate Alexa on our account. Authentication only needs to be done once per AVS app.

1. In the terminal, enter

sudo bash startsample.sh

2. You’ll see many lines scroll quickly on your terminal. Scroll up until you see: To authorize, browse to: 'https://amazon.com/us/code' and enter the code: [your unique 6-digit code here] 

3. In your Chromium browser, navigate to https://amazon.com/us/code 

4. Enter the provided code and click Continue. 

5. Click Allow. 

6. If you see “Success! Your registration is now complete,” you can start testing your AlexaPi. 

7. Say, “Alexa, good morning.” This should be a “hands-free” experience that does not require you to press additional keys to work. If your AlexaPi responds back, Great Job!

8. At this point, you can continue to test your AlexaPi by saying, “Alexa, ….” We have provided a list of “Fun Things to Ask Alexa” below.

9. In your Chromium browser, navigate to https://alexa.amazon.com

10. location. After you set your device location, you can listen to your current local weather by saying, “Alexa, tell me the weather.” 

11. To add Smart Home devices click on Smart Home on the left menu and enable account linking per the instructions from your device manufacturer.

12. When you are done with AlexaPi, return to the Terminal and press Ctrl-C. 

Fun Things to Ask Your Raspberry Pi-Powered Alexa 

Now that you have Alexa working on your Raspberry Pi, what will you ask her? Here are a few suggestions. 

• Play music.
• Can you sing?
• Talk Like a pirate.
• Can you rap?
• Open cat food. (Disclaimer: This is a skill that I wrote)
• Make me a sandwich.
• What’s the first rule of fight club?

Next Steps 

To mimic the lighted ring on the top of the Echo device, you could add a few LEDs to your project with this LED tutorial from Amazon.

If you have completed this tutorial and were disappointed that there was no Echo Show functionality, we will add screen functionality in a future article. 

Following up Apple’s new initiative to move its own Arm-based chips, Amazon yesterday said that the AWS data centers for its cloud-based Alexa and facial recognition services will also be moving from Nvidia processors to custom Amazon chips.

This marks the latest in a trend of major tech companies choosing to take their processing in-house. It’s also not the first time Amazon’s made this move, as most of its other AWS servers now run on the company’s Arm-based Graviton 2 chips.

Instead of Graviton 2, however, Alexa and the company's facial recognition service (named Rekognition) will now use the “Inferentia” chip, which focuses on speeding up machine learning tasks like translating text-to-speech and recognizing images.

According to Reuters, “Amazon said the shift to the Inferentia chip has resulted in 25% better latency...at a 30% lower cost.”

Currently, Nvidia supplies the chips for Alexa data centers, though Amazon hasn’t disclosed what chips Rekognition uses at the moment. Which raises an interesting question- we don’t know whether the Inferentia chip is Arm-based yet, but given that Nvidia recently bought Arm, that would mean Nvidia would still be providing the base Inferentia hardware.

Amazon currently hasn’t announced when we can expect new Inferentia-powered AWS services, but Rekognition might not be a top priority at the moment. The service has come under fire because of its use by law enforcement, and after the police killing of George Floyd this June, Amazon announced that police use of the tool would face a one-year moratorium.

Still, the shift to Inferentia is part of a new trend from major tech companies to ditch pre-developed chips in exchange for their own solutions. Currently, data center sales are Nvidia’s largest moneymaker, so we’re curious to see how the company will adapt.

Some gaming keyboards take the less is more approach, with slimmed down or low-profile designs that occupy up less desk space. The Thermaltake Level 20 isn’t one of those keyboards. Instead, this hefty clacker is extreme in all ways, from its size and weight to its RGB lighting and feature set. 

RGB is par for the course with the best gaming keyboards, but Thermaltake’s Level 20 (starts at $120 as of writing) takes things to the next level with and RGB lights running around three sides of the board and a light bar dividing the keys. Then Amazon Alexa steps in, allowing you to change RGB settings by speaking to an Amazon speaker, making this keyboard even more extra.

Thermaltake Level 20 Keyboard Specs 

 Design of Thermaltake Level 20 Keyboard 

The Level 20 almost looks like the classic IBM Model M keyboard’s wild cousin. Like the cult classic, Thermaltake’s keyboard is chunky with a vast amount of empty space above the FN row and a meticulously stamped logo. But unlike the Model M, which was obviously fit for the office with its muted color scheme, the Level 20 is loud, with more RGB than even a typical gaming keyboard.

To start, there’s a streak of LEDs running down the board, creating a glowstick-like barrier between the main keys and the nav and num keys. The bar consists of 5 individually controlled RGB zones, allowing for a multi-color effect. But at maximum brightness, all that flash can be distracting as it shines in your peripheral vision. South of the RGB bar is a stark cutout in the Level 20’s frame, which holds another RGB zone that’s a discrete treat for your eyes.

Contrastingly, I could almost never see the bottom RGB deck from my typical sitting position. The lights run through three sides of the keyboard but, ironically, barely go into the side of the board that faces you.  Sometimes I could see a peek of RGB in the corners, and if you’re gaming in the dark, the side lights will create a rave-ready glow on your desk. But otherwise the RGB here is virtually useless, which is saying something for RGB. I’d rather a hefty wrist rest to match the hefty keyboard than RGB zones that are respectively too distracting and too covert. 

Another (small) complaint is the appearance of the RGB in the light bar, deck and Thermaltake logo. These LEDs are shining through white plastic, which diffuses the colors slightly. For example, red looked like light red, and green appeared more playful and teal. This could interfere with your style. 

The Thermaltake Level 20 keyboard also makes its presence known through its size, measuring 19 x 7.3 x 1.7 inches and weighing 3.3 pounds. That’s heavier and more voluminous than your typical full-sized keyboard with media keys. For example, the Roccat Vulcan 122 Aimo is 18.2 x 9.3 x 1.3 inches and 2.5 pounds, and even the Corsair K95 RGB Platinum XT (18.3 x 6.7 x 1.4 inches, 2.9 pounds) will take up less space, despite it having a dedicated column of programmable “G”  keys. The Level 20 has feet with two levels of adjustments on its textured plastic base and rubber nubs, but you needn’t worry about this juggernaut of a keyboard sliding. 

You get all the media keys you need, including a rubbery and subtly clicky volume wheel and stiff buttons. But the Level 20’s bulk also comes from the aluminum top plate, which is 2mm thick and feels cool to the touch. Unlike the aluminum top plate on the Vulcan 122 Aimo, there’s no brushed finish on the Level 20, so it might not immediately look expensive. Instead, its dark, matte finish is a nice compliment to aggressive RGB engulfing the keyboard. For even more pizzazz, you can swap out 11 keycaps on the left side for the red, shimmery keycaps included in the box. Light bleeds through the red caps more, though.

The key layout offers many FN features, including the ability to tweak RGB’s effects, color and speed without launching software (or calling Alexa). You can easily launch up to 6 profiles with the help of the FN key and do macro recording, although there are no dedicated macro keys.

There’s a rare amount of USB passthrough with the Thermaltake Level 20 keyboard. Not only does it have a USB Type-A passthrough port, there’s also a 3.5mm headphone jack on the front edge of the keyboard. If you need your best gaming headset but your PC’s jack is hard to reach, this is great. The tradeoff is a hefty braided cable with a bulky three-way connector. If you want to use the keyboard and both of its passthrough ports simultaneously you’ll need a PC that has two Type-A ports and a headphone jack in decent proximity to each other (i.e. not on opposite sides of a laptop).  

 Alexa, Change My RGB

Alexa, why are you here? I admit that I never got aboard the voice assistant train. I don’t like the idea of placing internet-connected microphones around my home, and I’d much rather push buttons to control tech than speak commands (hello, I write keyboard reviews). But in a world of unreliable peripheral software and excessive RGB zones, some might appreciate the ease of changing their keyboard’s lighting by speaking to their Amazon Echo. 

Bringing Alexa control to the Level 20 seems easy enough. You enable “Alexa Skills TT RGB Plus” in your Amazon account, then log into Thermaltake’s iTake software. Afterwards, you can say “Alexa, tell Thermaltake” and then one of the supported phrases (see this PDF) to alter RGB colors, effects and speeds. You can do this through software and function controls too, but this gives you a way to bark commands at your keyboard. I don’t have any Alexa-powered devices in my house, and our New York lab is still locked down, so unfortunately I wasn’t able to see how well the Level 20 responds to Alexa commands. 

This feature may be more appealing if you own more products from Thermaltake’s TT RGB Plus line, which are all controllable via Alexa. With this feature available in water coolers, power supplies, RAM and more, you should be able to change the lighting throughout your system with a single voice command.

Typing Experience on Thermaltake Level 20 Keyboard 

Our review sample of the Thermaltake Level 20 came with Razer Green mechanical switches, but the keyboard’s also available with Cherry MX Blue or Cherry MX Speed switches. 

The Razer Green switch makes a clicky noise and has a tactile bump that you can feel as you travel down its 4.0mm total travel. It actuates at 1.9mm with 50 grams of force. Despite its name and color, the Razer Green is most similar to the Cherry MX Blue, which has the same amount of total travel but actuates at a further 2.2mm and with a much heavier 61 grams of force. The result was that they felt a little lighter on my fingers compared to Cherry MX Blues, and in a side-by-side comparison the Razer Green switches’ bump felt less prominent. 

The keycaps on the keyboard are your typical plastic, smooth without much in the way of texture besides the feeling of the font. Fans of Razer Green, Cherry MX Blue and other clicky mechanical switch types flourish in a chorus of clicks. But on the Level 20, the clicks are more like background singers for the louder and very slightly elevated keycaps. The keycaps make a hollow noise against the aluminum top plate, especially with the backspace button and space bar or when typing aggressively. The spacebar is also a bit shakier than I’d like, and the backspace key comes close to meeting that description too. Your experience may vary, particularly if you type very lightly. 

On the 10fastfingers.com test, I averaged 123 words per minute (wpm) with a 97.5% error rate. That’s speedier than my typical 110 wpm with a negligible dip in accuracy (I’m usually at 98%). The speed gain is likely due to the Razer Green switches requiring a lighter amount of force to actuate than Cherry MX Blues.

Gaming Experience on Thermaltake Level 20 Keyboard 

Your experience gaming with the Thermaltake Level 20 will depend on your switch choice. With our Razer Green switch-powered review unit, I certainly wasn’t fit for streaming or chatting unless I wanted to hear complaints about my loud clicks and keycaps.

Otherwise, the Razer Green requires slightly less travel to actuate (1.9mm versus 2.2mm) compared to the similar Cherry MX Blue switches. More noticeable to me, however, is how light the switches felt compared to MX Blues, which helped me bounce from key to key when gaming.

In Battlefield V, the Level 20 kept up with my soldier well, offering reliable input and the ability to quickly move on to the next action. However, the wobbliness of the shift key sometimes distracted me when I was holding it down to run. When I used the Level 20 with Overwatch, I experienced the same lighter sensation, and the keycaps were less of a distraction as I pressed different keys rapidly. However, those seeking a faster and more linear feel (where the keys go straight down without any bump along the way) should consider getting the Level 20 with Cherry MX Speed Silver switches, although that currently costs $20 more than our review unit. 

The Level 20 is supposed to have 100% anti-ghosting with full n-key rollover. It’s hard to press every key at once, but I did an informal test with an online tool and recorded 48 simultaneous inputs. 

Gamers will also appreciate on-the-fly macro recording, 6 onboard profiles and the thoughtful ability to switch the W,A, S, D for arrow keys, which is handy for the lefties. Another nice touch: FN and Scroll Lock activate Light Mode, which turnS on RGB for controls often used in games through 5 different configurations.

Features and Software on Thermaltake Level 20 Keyboard 

Amazon Alexa lets you control the Level 20’s RGB, but for much more customization you’ll need TT iTake software. It gives you the expected functionality, like lighting control and macro programming. 

Annoyingly, TT iTake is wonky and often crashed multiple times before I finished my adjustments. Once it crashed, I usually had to restart my computer to get it to properly launch again. Opening the software (or having it automatically launch at start) seems to deactivate the ability to control RGB on the keyboard, making customization frustrating. 

Because it has 4MB of onboard memory, you can create up to 6 different profiles that are launchable via 6 navigation keys and also importable.

If you can get through the app’s frequent crashing, you can use iTake to program every key, (except the media ones) to a macro function, media key, function key or launch software. There are also pre-made gaming profiles with commonly used functions suggested for 10 games, including CS:GO, League of Legends and Overwatch. 

There’s pey-key lighting, but Thermaltake starts you off with 11 RGB effects for the keyboard, and 12 for the RGB deck. There are two effects that take over both the keys and the deck. Music Mode reacts to music, and Temp Mode uses color coding to tell you your system’s temperature (one of six measurements) in Fahrenheit or Celsius.

The Level 20 also also works with Razer Synapse, which is a more reliable program and lets you sync the keyboard’s lighting with other RGB products that support Razer Chroma.

Thermaltake Level 20 Keyboard Versions 

If you like the Thermaltake Level 20 keyboard but don’t love it’s look, note that there are other versions available.

For about $140, there’s a version of this keyboard with a gray-colored top plate that makes it more apparent that it’s aluminum. You can find it with Cherry MX Blue or Cherry MX Speed Silver. Sorry, no Razer Greens here.

If you don’t need Alexa control or the RGB lightbar and sidelights, the Thermaltake Level 20 GT comes with Cherry MX Blue, Cherry MX Speed Silver or Razer Green switches, but as of writing we’re only seeing it readily available with Razer Green switches ($120). 

Bottom Line

The Thermaltake Level 20 has features no one really needs from their keyboard. We’re looking at you, Alexa. The RGB side lights and bars are a unique touch, but one of them you’ll hardly ever see and the other can be distracting when gaming or working.

With options for linear or clicky Razer or Cherry switches, you can find your preferred feel for gaming, and the Level 20 is clearly made for gamers, with features like macro recording and onboard profiles. 

If you’re overwhelmed by the RGB, consider the Thermaltake Level 20 GT, which is slightly cheaper, but as of writing we’re only seeing it with the Cherry MX Silver Speed switches. Another option is the Patriot Viper V765, which also has side RGB lights and is cheaper at $100.

But if you’re looking for a fearless design and perhaps live in an Alexa-friendly home, the Level 20 is ready to listen to your voice commands as you battle and/or type. 

There are so many ways to make your home into a smart home using a Raspberry Pi. We've covered home DIY Pi projects that make your home smarter before—from garage door openers to AR light switches. But this latest project, from a maker known as Rleekc, brings the smart home concept outside by using a Raspberry Pi to control lights in the backyard.

The backyard lights use a relay switch that can receive input from a 433 MHz remote control. A 433 MHz receiver and transmitter modules are necessary for the Raspberry Pi to interact with the relay. The Raspberry Pi runs a Flask-based web server that's used to adjust the lights. 

The remote has a limited range of about 50 feet, but Rleekc was able to copy the remote signal to be retransmitted with the 433 MHz modules. This made it possible to control the lights through the Pi server.

The biggest challenge with this project is interference. Many devices, like TV remotes and garage door openers, operate on the same wavelength. To alleviate this issue, Rleekc invested in a faraday blanket to block radio waves and isolate the signal.

You can explore the original post on Reddit and check out the official project write-up on Github. Be sure to follow Rleekc for more updates and future Raspberry Pi projects. 

Security researchers from the University of Michigan and the University of Electro-Communications (Tokyo) have invented a new technique called “Light Commands” that can take over voice assistants, including Amazon’s Alexa, Google Assistant, Facebook Portal and Apple’s Siri, through the use of a laser beam. 

What Are Light Commands?

According to the researchers’ website, Light Commands is a vulnerability in Micro-Electro-Mechanical Systems (MEMS) microphones that allows attackers to remotely inject inaudible commands into popular voice assistants. The researchers were able to use this vulnerability to deliver an inaudible, invisible command via a laser beam targeted at the voice assistants. 

The technique requires the laser to actually hit the target device’s microphone port, which could get significantly more difficult as the distance gets larger. However, the researchers also noted that the attackers could use binoculars or a telescope.

The longest distance they tested was 110 meters (360.9 feet), which was the length of the longest hallway they could find at the time. But the researchers noted that Light Commands could easily travel much longer distances. 

In addition to voice assistants, the researchers said the attack works against other devices that use MEMS microphones and could be targeted by a laser beam -- including smart speakers, tablets, smartphones -- across large distances. 

What these researchers have demonstrated is that security attacks are getting increasingly advanced and that companies need to consider not just software attacks but also physical attacks against their devices. Many companies haven't even caught up on properly securing their device's software yet, so it remains to be seen what they'll do about their devices' physical security.

Potential Mitigations

Voice recognition reportedly can’t protect users from this flaw, -- even if they switched it from being off by default -- according to the researchers, because it could also be easily bypassed with text-to-speech technologies that can be made to imitate the owner’s voice.

The researchers proposed a few mitigations that could work better against this sort of attack. 

One requires the user to authenticate with each new command, such as by answering a random question (something akin to a voice-based version of a CAPTCHA test).

Another requires manufacturers to increase the number of microphones on the device that are needed to receive a command. The laser beam can only target one of microphone at a time, so requiring two mics to listen to each command would, in theory, prevent this attack. 

The final mitigation recommended is to add a physical barrier above the microphone's diaphragm to prevent light from hitting the microphone directly. 

People expect their virtual assistants to be discrete, which is why devices featuring Alexa, Google Home and other assistive technologies are supposed to make it clear when they're recording. But ZDNet reported yesterday that hackers can still exploit vulnerabilities in Alexa and Google Home to use them as covert surveillance tools. 

The exploits in question simply require app developers to release malicious updates that insert long sequences of the "�" character in responses to user queries, allowing for a long duration of silence, whereby the devices read that character and their microphones remain active. Someone could then use those microphones to eavesdrop on Alexa and Google Home users.

Attackers could go one step further by having their malicious app impersonate Amazon or Google and tell people to share their account credentials so they can sign back into their services. Neither device would do that normally, but unsuspecting users might not think twice about compromising their account security. 

These devices technically disclose their active microphones with a status light, so observant people could notice that something's up. But it's also very likely for people to miss or ignore the light entirely. Part of these devices' appeal is the ability to shout commands across rooms without interrupting the task at hand; having to stare at a status light doesn't jive with the reason people use these products.

Researchers disclosed similar vulnerabilities in Alexa and Google Home throughout 2018. Security company SRLabs discovered their continued presence earlier this year. The company reportedly told Amazon and Google about the issue, but it remains unresolved, even though SRLabs told ZDNet that "finding and banning unexpected behavior such as long pauses should be relatively straight-forward." 

Here's how SRLabs said Amazon and Google should address the issue in a blog post:

"To prevent ‘Smart Spies’ attacks, Amazon and Google need to implement better protection, starting with a more thorough review process of third-party Skills and Actions made available in their voice app stores. The voice app review needs to check explicitly for copies of built-in intents. Unpronounceable characters like '�,' and silent SSML messages should be removed to prevent arbitrary long pauses in the speakers’ output. Suspicious output texts including “password“ deserve particular attention or should be disallowed completely."

The company also shared four videos on its YouTube channel showing the exploits for conducting phishing attacks and eavesdropping on Alexa and Google Home users. 

For its part, Google told ZDNet that it "removed the Actions that we found from these researchers "and is "putting additional mechanisms in place to prevent these issues from occurring in the future." Amazon has remained silent.

Editor’s note: This article was originally published on February 1, 2019. It was updated in September to reflect software updates and features added since and the Focals Showroom app. 

Smart glasses have had a tough time going mainstream. Even with the backing of major brands, attempts like Google Glass have failed to open up the market, partially due to their alienating looks. Focals smart glasses from North address this by looking, fitting and working like traditional prescription glasses (prescription lenses start shipping spring this year). Plus, your tech addiction is your little secret with an augmented reality (AR) display only the wearer can see. The commitment to stealth continues with Loop, a plastic ring that lets you control the glasses discreetly through its five-way joystick.

Focals are currently only available after two in-person fittings (for more on North’s detailed fitting process, see our first hands-on with Focals) in their Brooklyn, New York or Toronto, Canada stores or via the Focals Showroom app for iPhone X or newer. It's tempting as Focals cross a huge smart glasses barrier by offering functionality in a form that stands a good (but not perfect) chance of passing for regular glasses. However, while we enjoy apps like Spotify and Weather, better image quality would make the $599 / $799 CAD price tag more forgivable.

Focals Smart Glasses and Loop Ring Specs

Focals use a Qualcomm APQ8009w system-on-a-chip (SoC), which runs on four Arm Cortex A7 CPU cores at a clock speed of up to 1.09GHz. The SoC is marketed for smartwatches, with features like Bluetooth and WiFi connectivity and a Qualcomm Adreno 304 GPU.

North designed Focals with IP55 certification in mind. This standard calls for dust-resistance and the ability to withstand water from a nozzle. That means Focals should endure splashes without breaking. North doesn’t recommend wearing them in the shower or pool.

The Loop is made for IP66 certification and is therefore more dust- and water-resistant than the glasses. That specification calls for the ability to handle “water from heavy seas or water projected from jets.” North says the Loop is suitable for washing your hands but not swimming. It’s available in 10 sizes ranging from 6 to 15.

Design: Stealthy Yet Stylish

Focals are by far the most fashionable smart glasses I’ve seen. Many smart glasses and mixed reality (MR) headsets have tried to improve their appearance, and public perception, by cutting size down and bringing in more traditional designs. Some, like the Vuzix Blade and nreal light, reduce the computer-on-your face look by mirroring sunglass designs.

Focals, however, are in their own category, looking like regular prescription glasses and come in various styles: classic shape in black, or for $100 extra tortoise or gray fade (our review sample is classic in tortoise). Round-style frames are "coming soon," according to North. Focals could very nearly pass for regular glasses if it wasn’t for their chunky arms and the small projector near the right lens that’s visible to people looking at you.

Focals are primarily made of die-cast aluminum and Swiss Grilamid, which is supposed to be a strong and flexible -- yet shape-retaining -- polymer. That’s likely why Focals are so bendy. I can hold them by their arms and twist them disturbing the lens alignment or tug on the arms and pull them pretty far apart. The material is also supposed to make them more comfortable long-term.

Copper accents connect the rubber nose pads, and both arms fold in half thanks to stainless steel springs.

Speaking of those arms, the halves closest to the temples are quite thick. At their thickest point they’re about twice the width of those of my regular glasses. This isn’t an automatic tell that you have a computing machine on your face, but Focals do look chunkier than your average pair of glasses.

The left arm of my review sample is bare black on the outside, while the inside subtly reads “Focals by NORTH” near the temple and “CLASS 1 LASER PRODUCT” near the tip. The right arm is also bare on the outside.

The inside, however, holds the holographic display projector, which uses a display technology called retinal projection to project photons, or light, or raster graphics, onto the retina. When the projector is activated, it’s not visible from the outside. It projects images onto the right eye only. This advanced retina display also calls for precise measurements in the aforementioned fitting process.

On the bottom of the right arm is a small square area for connecting the charger, a small speaker/microphone and the power button.

The Loop controller ring comes in black and is mostly made of polycarbonate with gold-plated charging contacts. You’ll hear a clicking noise in the glasses every time you use the joystick, unless you turn all sound off.

Both the Focals and Loop charge via the included case, which has a USB-C port in the back for plugging in the included power adapter. The adapter comes in smooth gray to match the case, and can also serve as a portable charger for the devices. The case, a smooth, soft, felt-like gray box, is larger than a traditional glasses case. Its handy internal lights illuminate when your devices are charging.

Focals don’t just look like prescription glasses, they're also available with prescription lenses (-4 to +2SPH and 0 to -2CYL) for an extra $200. North won’t offer eye exams, so you’ll have to know your prescription beforehand. This is a bit of a drawback because that means one more appointment and/or phone call if you don’t know your prescription offhand (more on availability and insurance in the Availability section below).

My Most Comfortable Glasses Ever

I wore my Focals and Loop around California for a week, meeting with friends, getting dinner and going to the beach. They stayed snug on my face, even more so than my real glasses, which tend to slide down my nose. North has a two-part fitting process, which includes a 3D face scan and fitting, followed by some final physical tweaking, such as on the nose pads, when you pick them up. This, indeed, led to a pair of specs that fit perfectly, perhaps my most comfortable specs ever. This is also helpful in keeping the AR visible on command without needing to adjust the glasses.

Weighing less than a fourth of a pound, Focals didn’t feel heavy on my face. They are certainly the most lightweight and cozy XR glasses I’ve tried yet.

For even more versatility, North includes a pair of clip-on sunglass lenses with each Focals. They’re available in black with gold or black rims and claim to block 100 percent of UVA and UVB rays.

And the Loop, while a bit chunky and plastic-y for my taste and not the jewelry style I’d typically buy, did not bother my hands long-term. However, I did occasionally push the joystick in unintentionally when doing common tasks like picking things up from the ground.

Focals Apps

Focals wearers must download North’s free companion app on iOS or Android. The app is simple and easy-to-use. It displays the battery status of the Focals and Loop and allows you to calibrate the Focal’s display. It’s also where you log into Amazon for Alexa and Uber, store up to seven contacts for easy texting and two locations for easy directions or Ubers and contact support.

Note that your smartphone also has to connect to the glasses via Bluetooth at all times and allow location tracking and numerous other privacy permissions. North says Focals and Loop can work from up to 9 feet (2.7m) away from your phone.

Home Screen and Notifications

The Focals home screen comes up whenever you press the Loop’s joystick in. It’s a colorful display of the time and date and also has options for adjusting the volume to four settings, including off, toggling sound on or off (a little redundant) and switching between three brightness settings. If you have a meeting coming up soon in your smartphone’s calendar, a reminder shows here underneath the time.

Scrolling to the left brings the notifications screen, which displays any new notifications coming to your phone, including texts, emails, reminders and anything else that would pop up on your phone. Typically, you can read the first few lines of an incoming text or email here; however, you can’t click on any links or view media. There’s an option to clear all notifications at once.

As far as usefulness goes, this was quite handy on the go. I was able to do things like read messages while online for lunch, holding my wallet in one hand and a water bottle in the other. I was able to determine if incoming notifications were important enough to put down my food (they rarely were).

Spotify

Spotify control is one of the best features of the Focals, especially for a daily rider of the New York City’s overstuffed subways. Without having to free or raise your hands, I could easily play/pause and change tracks on Spotify with the Loop ring (up to play/pause, left and right to change songs). If you’re You can even do this without the Focals’ display turning on, saving you battery. Moving the Loop’s joystick down is also supposed to let you control volume, but I couldn’t get this to work with a Samsung Galaxy S10. North’s reliable in-app support (or luck) should help you fare better.

Also disappointing was that the Focals’ Music ID feature, which pretty much works like Shazam and is supposed to identify the song that’s playing aloud, wouldn’t work for me either. I tried with three songs, Michal FK’s “Pulsar,” Kendrick Lamar’s “DNA.” and a song everyone should know, Britney Spears’ “... Baby One More Time.” No luck. All I got was a “Thinking” message with no results.

Health and Fitness

Focals take a page out of one of the most successful modern-day wearables, smartwatches, by incorporating health-related apps, Google Fit and, if you have an Android phone, Digital Health for screen time management.

With Google Fit, there’s plenty of reason to consider Focals as your smart health device of choice. Google Fit includes a step counter, improvements over the previous week and a calorie tracker.

Digital Health caters to growing focus on managing time spent on your phone. Interestingly, the Digital Health app doesn’t count Focals time as screen time, so you can cheat by using your Focals instead of your smartphone, encouraging users to use and depend on their Focals more (handy for North).

There’s also a feature that reminds you to drink water and lets you mark down how much you’ve water you’ve had. It sounds stupid, and I don’t deny that it is. However, it did come in handy for me once when I went downstairs for a glass of water, got distracted by something else and was on my way back up the stairs sans hydration. The pop-up coincidentally arrived just in time to save me a wasted trip down and up the stairs. Everyone should already know to drink water regularly, but if this habit is new for you, this feature could help.

Productivity

With support for Slack, EverNote and OneNote, Focals can help you get work done. Pop-up notifications and pre-written responses or speech-to-text responses for Slack mean it’s easy to do things like answer your boss on a lunch break via Slack or read a document while heading to a meeting with the glasses. However, with the clarity issues I experienced, I won’t be reading anything longer than an email on my Focals.

The Focals Rewind feature lets you record and save 30 seconds of a conversation. It’s creepy, since there’s no light on the Loop or glasses to let people know you’re recording. ANd there’s not much use to it with the recording time limited to 30 seconds. Additionally, this only works if you’re currently in a meeting that you logged in your calendar synced with Focals. And you can’t get the recording until the meting is over, according to your calendar. At least the recording sounded… JUST the words “that’s good” not even a recording.

Fun and Games

In the Maurice game, you make a cow jump on top of colored logs one-by-one by pushing in the Loop’s joystick. The game is simple -- yet colorful and potentially addictive -- and a sign of other games Focals could support in the future.

However, the Focals got warm after just a short time playing this simple game. After 10 minutes, the right arm felt warm to the touch. While it was a little warm on my face, it wasn't so hot that I needed to take them off.

Focals’ trivia game challenges with 20 new multiple-choice questions daily (at the time of writing it was The Office-themed). You can see the average score (3/20 today) and your score.

You can also caught up on the latest MLB, NBA or NHL games with real-time updates and scores with your favorite teams selected. If you’re a diehard fan who needs to see the play-by-play, your favorite smartphone app is still necessary, but for quickly checking the score during a date without getting scolded, this is great.

Amazon Alexa

Alexa is always at your beck and call with Focals. It appears any time you hold in the joystick, so it’s pretty easy to accidentally summon the voice assistant.

Alexa on Focals does what you’d expect. It can answer questions, tell jokes and list nearby restaurants, speaking via the speaker on the right arm. However, it loses some capabilities on Focals, like music and radio playback or the ability to bring up images or videos.

But Alexa also works in conjunction with Focals’ other apps. For example, I asked Alexa where the closest mall is, saw a list displayed, pushed the Loop’s joystick in on the closest one and saw an option for directions or calling an Uber to go there.

Alexa listens pretty well via Focals. I never had a problem with the TV or other background noises. However, it usually took at least 30 seconds to respond, sometimes more. And sometimes it straight up ignored me, with the blue circle flashing for 30 or more seconds before disappearing without any result or error notification.

Uber

You can call an Uber on Focals by selecting it the option from one of the two locations you’re allowed to store on the Places section in the smartphone app. If you’re going somewhere else, you can go the Places menu in the Focals and use voice command to write out an address.

Regardless of how you call your Uber, Focals will keep you up to date, displaying updates like when your car is coming and your ETA.

Texting

You can text with Focals using pre-written responses or voice-to-text by either responding to an incoming message or starting a new text to one of seven contacts you can favorite in the Focals smartphone app. You can also read your texting history, up to 12 incoming and outgoing messages, on Focals. Again, you can’t click on links or read images. Instead, it’ll say “This message has media. Check it on your phone.”

But there’s a pesky caveat for iOS users like me: Due to some restrictions Apple put in place in protection of its messaging app, Focals sends all texts from a different phone number. When texting someone with Focals for the first time, it’d first automatically send a text (from a different number than mine) saying, “Hey it’s Scharon Harding. I’m messaging you from my Focals by North.” It’d also start a new thread in my phone. My phone would go off every time someone texted me and each time I sent a text with Focals, as if Focals was texting my phone as well. This led to my phone going off at least twice as much and a messier inbox.  Android users won’t have this problem because Focals can sync directly with the Android messaging app.

This behavior also brought a few interesting responses, as you can see. Among others, my ever-cautious father refused to answer any texts from this strange Focals number. I wasn’t surprised. Besides the number being unknown, thanks to autocorrect I could never text the words “This is Scharon”—only “This is Sharon” or “This is s c h a r o n.”

I also struggled a tad with voice-to-text. Words crawled onto the screen much slower than I speak. It autocorrected in front of me, then provided an option to send when it thought I was done speaking. But sometimes I have a lot to say in texts and don’t know what it’s all going to be when I start. And watching the Focals write out and edit my words threw me off.

Further, Focals’ display automatically goes to sleep after about 8 seconds to prevent the AR from becoming intrusive and preserve battery life. So sometimes, if I paused too long in between words the display turned off and I’d have to start my text over. This often seemed to happen in fewer than 8 seconds though, forcing me to restart my messages plenty of times.

While it was not flawless, I was still impressed with the Focals’ listening comprehension here and also got a kick out of how Focals paired incoming message alerts with an emoji based on the messages’ contents. I appreciated seeing an unimpressed smiley face next to some of my mother’s texts.

If you don’t mind the limitations of pre-written responses or the finicky behavior of voice-to-text technology, Focals’ texting app can certainly help with multitasking.

Calendar

Focals sync with your phone’s calendar to display upcoming events for the day. You can only see today’s schedule in the Focals. You can’t click on an address in an event or add new appointments through the smart glasses.

Directions

Focals has turn-by-turn and compass-based directions, powered by Mapbox customized maps, in the Places menu. You can store two addresses that’ll always show in Places, allowing for quick directions or an Uber, in the smartphone app. The Places menu also automatically displays one nearby point of interest. Alternatively, you can enter an address with voice command. Directions are not recommended for use while driving.

Helpful for city folk, the Focals also tell you where the closest public transportation location is. My Focals told me that the next buses to arrive outside my apartment were 5 and 10 minutes away. However, like many directions apps, it didn’t consider transportation lines further away, like the subway station 0.8 miles from my apartment that I walk to daily. So for the most detailed depiction of your travel options, you still can’t beat looking at a map or street smarts.

If you save your work or home address into the Focals app, the Focals’ home screen will also tell you the quickest way to get there via public transportation during rush hour. The smart glasses also let you send an ETA message based on the route by clicking on it. This is all potentially useful for those constantly rushing their commute or with multiple stops to make. But it’s forgettable for those who routine treks in and out of work daily.

Weather

The Weather app shows you the current weather in your location. Scrolling down shows you the current weather, plus three more, “Aft,” “Eve,” “Night” and “Morn” (tomorrow), depending on the current time of day. You can’t check the weather for any other day or by hour.

Like most weather apps I’ve tried, Focals’ Weather app doesn’t seem completely accurate. At one point, I could see it was raining, and even Alexa confirmed this, but Focal claimed mere “fog.”

Battery Status

By scrolling all the the way to the right, you can see the battery status of the Focals and Loop at any time, shown in individual percentages.

Image Quality

The most impressive part of Focals’ is that no one will know when you’re using apps, since the AR display is only visible to the wearer.

It works by creating red, green and blue light that is manipulated to make text and images the Focals’ projector sends out. Next, there’s a holographic lens in the right eye that’s embedded with a transparent film designed to interact with red, green and blue wavelengths only. Everything else passes through. According to North, “when our specific wavelengths of light hit the transparent film, it acts like a mirror and bounces the light back towards your eye placing the image directly in your line of sight where only you can see it.”

The first thing I noticed about Focals’ display is the lively colors. Like an ice cream parlor, the AR is filled with bright pinks, blues, yellows and green. Never a dull notification here. It makes things even more fun by adding emojis for incoming notifications. Even activating the display, which takes about a second, yields a mild animation, with the time climbing up in gradient effect and color.

In the sun, at night, or with the sunglass clip, the image remained bright and bold, but I did occasionally see ghost images, or double vision. For me, this happened occasionally indoors near sunny windows. I’d see a light green reflection of the display off to the right side. North told me this is a result of green light bouncing off a lens layer meant for red or blue light and that it tends to happen in darker environments. Turning brightness down to the minimum usually eliminated the effect.

The biggest problem, though, was the flicker effect I experienced. Images always appeared to be moving very slightly, similar to a CRT monitor. This hurt the clarity of the display, which was still legible but a touch difficult to look at for more than a couple seconds, especially for text. North told me this was likely caused by my glasses’ refresh rate being too low. After receiving a replacement pair, the image seemed sharper but at times still looked like it was quivering a tad. It’s possible I’m more sensitive to the effect than others.

Image quality stayed pretty consistent in different environments, except for when I looked directly out of a sunny window (without the sunclips on). With that much light hitting my eyes, some text on the left side faded in and out of visibility when reading text messages or the day’s weather.

Battery Life

North claims Focals last up to 18 hours and the Loop up to three days with “intermittent use.” As mentioned, you can always check the status of the Focals and Loop on the companion app or by scrolling all the way to the right on the smart glasses.

I used the Focals in lieu of my phone as often as possible — which meant a lot of texting and reading notifications, checking the weather and directions at least once and playing a game for at least 5 minutes—at max brightness and volume. They lasted 7-10 hours a day, while the Loop survived a smidge over 3 days before needing a recharge.

As mentioned earlier, the Focals and Loop’s case is also their charger. They take up to 2 hours to fully charge simultaneously. The case also fully charges in four hours, for use as a portable charger. A fully charged battery case is supposed to last for three full charges, but mine never lasted for 2 full charges of the Focals only (not the Loop). I wouldn’t recommend going on a weekend trip without the power adapter.

Heat

After 2 minutes of continuous use the Focals were warm to the touch. After 5 minutes, I felt a gentle warmth upon my face. Again, this wasn’t enough heat to make me take the glasses off but was still bothersome. Most of the warmth was on the right arm, especially on the top (the part that doesn't touch the face) by the temple.

After playing with the Focals for 5 continuous minutes at maximum brightness and volume, the right arm’s hottest point was 44 degrees Celsius (111.2 degrees Fahrenheit).

Availability

Focals can be bought in person via two fittings in their Brooklyn or Toronto stores. This ensures the glasses fit your face and the AR is properly positioned for your eye. North also offers pop-up showrooms for these services. If you can't make it to one of those locations, you can also get fitted for and buy Focals off North's Showroom app; however, that's currently only available for newer iOS phones (iPhone X or newer). 

Focals cost $999 without prescription lenses. Prescription lenses adds $200 to the price tag. Those who get prescriptions, however, will be able to use their insurance to help pay, North told me. Pricing will depend on the customer’s insurance provider/plan.

Bottom Line

Focals are worth a look if you’re ready for smart glasses. With AR only you can see, they're discrete, and there’s a good chance people will mistake them for regular glasses or, with the included-clip on lenses, sunglasses.

But a lot of what you’re paying for is that look. In terms of apps, Focals has the minimum, but I’d like to see more. Control over music, a video streaming app, a voice-to-notes taker or even some games would make the $999 specs more appealing. While that price is in line with competitors, you can still get a laptop for that amount, so the more functionality, the better.

Among Focals’ current apps, there are some small, but noticeable, limitations, such as iPhone users’ texts sending from a different number and the inability to click on links or media or see the calendar or weather for upcoming days. The glasses usually alert you when you have an incoming/missed call, but a couple of times my review unit failed to do so. 

The image quality also has room for improvement. In addition to the occasional ghost images, two of my review samples displayed content that was legible but looked slightly shaky, like a CRT monitor, making text in particular a little irritating to look at with mild, unwanted blur.

But if you’ve been waiting for a pair of smart glasses that offer function without obliterating your fashion cred, there’s no better choice currently than Focals. With the Vuzix Blade currently starting at $799, the Focals are $200 cheaper and much better-looking. However, the Blade’s display is much sharper. The mandatory Loop controller can be viewed as a burden, but it’s a small price to pay for subtlety. If you’re willing to pay the premium price required of smart glasses and can get to Brooklyn, Toronto or a newer iPhone, Focals promise the best style, a fun, easy-to-use (but basic) user experience and premium fitting.

Plain Jane SATA

Lexar’s got a bunch of new SSDs in the market, but are they any good? Today, we check out their NS200, a SATA 6 GBps SSD that promises to deliver speeds of up to 550 / 510 MBps read/write and the ability to keep on chugging through all your applications with fairly high endurance capabilities. And, to top things off, the device is priced competitively, but it won’t deliver you all the value alternatives will as it lacks support software at this time, at least in the US.

Lexar, an old name in the storage industry known best for the company’s Jump drives is now the consumer brand for the Chinese company Longsys. With the recent adoption, the company now has a plethora of SATA and NVMe SSD products for anyone looking to upgrade their storage to flash. The NS200 is the company's fastest SATA SSD and after looking at the design, it becomes apparent why.

Rather than go the DRAMless route, Lexar’s NS200 boasts a Silicon Motion SM2258 SATA controller, which utilizes a DRAM cache, and Micron’s 64L TLC NAND. These are high-class components that have proven themselves in other SSDs like the ADATA SU900, Crucial BX300, and most notably the Crucial MX500. But, that’s not all. The NS200 is a SATA 6 GBps SSD that is available in three mainstream capacities: 240GB, 480GB, and 960GB. Based on these capacities, the company has also opted to increase the over-provisioning for added performance and reliability over usable space for the end-user, something their lower-tiered NS100 doesn’t.

Specifications

The NS200 is capable of up to 550 MBps read and 520 MBps write speeds under sequential workloads and in terms of random 4K performance, Lexar rates it at up to 95,000 / 90,000 IOPS read/write.  While faster than an HDD, this is significantly lower than what most NVMe SSDs are capable of today. Fortunately, the NS200 is priced fairly competitively at the larger capacities. The 480GB and 960GB NS200 are about $0.11-0.13 per GB while the 240GB model is priced at about $0.19 per GB.

Lexar’s NS200 supports a bunch of standard features such as S.M.A.R.T. data reporting, Trim, and secure erase support. Its endurance ratings are impressive for a SATA device. They aren’t as high as Samsung’s 860 EVO, but with LDPC ECC and some extra over-provisioning, figures come in at up to 480TBW for the 960GB model, surpassing that of the WD Blue 3D and even Crucial’s MX500. What it lacks, however, is a 5-year warranty, rather it features just a 3-year warranty like most cheap SSDs.

Software and Accessories

Lexar includes their own SSD Toolbox for download on their website, Lexar SSD Dash. It provides up-to-date drive health, drive condition, and S.M.A.R.T status information to help you monitor the device.  However, unlike most manufacturer’s they do not offer any free cloning software. As well, at the time of writing, the download for their software is unavailable in the US due to government requirements blocking it.

A Closer Look

Lexar’s NS200 comes in a 2.5-inch form factor and is 7mm thick, like most other 2.5-inch SSDs. Looks-wise, the design is clean and well branded, it won’t take away from aesthetics like WD’s Blue 3D can, but the case is all plastic. Thus, it is rather easy to twist and bend, unlike Samsung’s 860 series and Crucial’s MX500 SSDs which feature a full metal case. Not a huge deal-breaker, but it’s not on par feel wise.

Within the case is a small, half-sized PCB with four of Micron’s 64L 3D TLC NAND flash chips that have been labeled as Lexar’s own. Managing the flash, we see Silicon Motion’s SM2258 SATA controller, one that features a DRAM cache for accelerating access to the FTL map, which helps to keep up performance, even under strenuous workloads. For this task Lexar has opted to utilize a bit of SK Hynix’s DDR3 memory, which is more than fast enough.

MORE: Best SSDs

MORE: How We Test HDDs And SSDs

MORE: All SSD Content

1TB Performance Results

Comparison Products

We pitted the Lexar NS200 against its near-identical twin and one of the highest-value based SSDs in the market, the Crucial MX500. As well, we threw Samsung’s premium-cost 860 EVO into the mix and Intel’s 545s with the slightly newer SM2259 controller powering it. We also threw in WD’s Blue 3D SSD and Crucial’s DRAMless BX500 which utilizes Silicon Motion’s SM2258XT controller. Additionally, all feature 3D TLC NAND flash, except for one. We also included Intel’s 1TB 660p for our application comparisons as we lack 500GB sample to compare directly. It comes with Intel’s 3D QLC NAND and is powered by an SMI SM2263 NVMe controller.

Game Scene Loading - Final Fantasy XIV

The Final Fantasy XIV StormBlood benchmark is a free real-world game benchmark that easily and accurately compares game load times without the inaccuracy of using a stopwatch.

Lexar’s NS200 display’s some very impressive game load performance. With a total load time of 21.99 seconds, the NS200 powers to second place in our comparison pool. We can also see that it is significantly faster than an HDD and not too much slower than Intel’s 660p.

Transfer Rates – DiskBench

We use the DiskBench storage benchmarking tool to test file transfer performance with our own custom 50GB block of data. Our data set includes 31,227 files of various types, like pictures, PDFs, and videos. We copy the files to a new folder and then follow up with a reading test of a newly-written 6.5 GB file.

Again, the NS200 powers to second place, displacing the more expensive Intel SSD 545s and Samsung 860 EVO with a data rate that is 17 MBps faster. Compared to the low-end BX500 and WD Blue HDD, Lexar’s NS200 is about three times faster at file copies. Read performance ties Samsung and Intel too, with a rate of 503 MBps SATA performance doesn’t get much better than this. Still, Intel’s 660p takes the win, outperforming the whole group with its NVMe goodness.

Trace Testing – PCMark 8 Storage Test 2.0

PCMark 8 is a trace-based benchmark that uses Microsoft Office, Adobe Creative Suite, World of Warcraft, and Battlefield 3 to measure the performance of storage devices in real-world scenarios.

Tasking the drive with some applications we can see that the NS200 isn’t quite as fast as the competitors and it falls into 5^th place here. It is almost as fast as the MX500 in this instance, but it can’t keep up with the likes of the Samsung 860 EVO if you are looking for one of the best performing SATA SSDs out. It is still light-years ahead of a standard HDD, however.

SYSmark 2014 SE

Like PCMark, SYSmark uses real applications to measure system performance. SYSmark takes things much further, however. It utilizes fourteen different applications to run real workloads with real data sets to measure how overall system performance impacts the user experience. BAPCo's SYSmark 2014 SE installs a full suite of applications for its tests, which includes Microsoft Office, Google Chrome, Corel WinZip, several Adobe software applications, and GIMP. That also makes it a great test to measure the amount of time it takes to install widely-used programs after you install a fresh operating system.

Lexar’s NS200 was able to install the software suite at a rapid pace, faster than any other SSD in our comparison pool. After running the benchmark, the NS200 attained a responsiveness score of 1607 points, which nearly matches the Crucial MX500, but again, Samsung and Intel are in the lead with slightly better performance.

Synthetic Testing - ATTO

ATTO is a simple and free application that SSD vendors commonly use to assign sequential performance specifications to their products. It also gives us insight into how the device handles different file sizes.

Sequential performance falls in line with the competition across the various file sizes tested. In ATTO the NS200 maxes out at 564 / 527 MB/s read/write.

Synthetic Testing - CrystalDiskMark

CrystalDiskMark (CDM) is a simple and easy to use file size benchmarking tool.

Like its performance in ATTO, Lexar’s NS200 delivers sequential performance of over 560 /520 MBps read/write, which is similar to the competition. Its 4K random performance is a little weaker, however. At a QD of 1, it delivers 34 MBps read and 106 MBps write. While write performance is on par with the rest, read performance lags a bit which is why we saw slightly lower application performance in our test suites. Still, it is much better than the Crucial’s BX500’s scores, especially when averaging out QDs 1-4.

Sustained Sequential Write Performance

Official write specifications are only part of the performance picture. Most SSD makers implement a pseudo-SLC cache buffer, which is a fast area of SLC-programmed flash that absorbs incoming data. Sustained write speeds can suffer tremendously once the workload spills outside of the pSLC cache and into the "native" TLC or QLC flash. We use iometer to hammer the SSD with sequential writes for 15 minutes to measure both the size of the pSLC buffer and performance after the buffer is saturated.

Like almost every SSD out these days, the NS200 features a pSLC cache. Fortunately, this cache is large and dynamic. After writing 143GB of data, the buffer filled and sustained write performance degraded to an average of 275 MBps. Because the cache is so large and dynamic, sustained performance takes a hit compared to that of the MX500 or other competitors which feature smaller caches for improved consistency. Still, this is plenty of performance for most home users that task the drive small write bursts throughout the day rather than large sequential transfers.

Power Consumption

We use the Quarch HD Programmable Power Module to gain a deeper understanding of power characteristics. Idle power consumption is a very important aspect to consider, especially if you're looking for a new drive for your laptop.

Some SSDs can consume watts of power at idle while better-suited ones sip just milliwatts. Average workload power consumption and max consumption are two other aspects of power consumption, but performance-per-watt is more important. A drive might consume more power during any given workload, but accomplishing a task faster allows the drive to drop into an idle state faster, which ultimately saves power.

The NS200’s power consumption is very well managed. Averaging 1.63W and hitting a max of just 3.31W during our 50GB transfer, it scores the highest on our efficiency chart. It also consumes just 51mW when at idle with Link Power Management enabled. That number grows to under half a watt when disabled, but it still beats the MX500’s consumption results.

MORE: Best SSDs

MORE: How We Test HDDs And SSDs

MORE: All SSD Content

Conclusion

Lexar’s NS200 is a decent SSD for those looking to upgrade from a hard drive or a smaller capacity SATA SSD. It speeds up your boot and application load times as well as will hold you over while fragging noobs in your favorite games. But it lacks in overall value.

Performance is nearly on par with most mainstream competitors and it even outperformed our comparison pool during game loads and file transfers. It lags slightly in random read performance though. Fortunately, in day-to-day use cases, this difference goes almost unnoticed. And, sequential performance is good enough to keep up with those who work with various 10—50GB files and then some. Although, competitors do offer slightly better-sustained speeds once their pSLC caches fill.

Unfortunately, it doesn’t come with any cloning software and, at the time of writing, the U.S. government prohibits U.S. customers from downloading their SSD toolbox, Lexar Dash. As an entry-level device, it also lacks a 5-year warranty that most mainstream competitors feature. But it does boast some very good endurance figures that outclass even the Crucial MX500 and WD Blue 3D.

Pricing is in line with most SATA SSDs these days, but we recommend waiting for it to go on sale. At almost $0.13 per GB, the 960GB model is one of the best values, but the 480GB model we tested for this review still provides pretty good bang for your buck. We would recommend the Crucial MX500 over it, however, for a number of reasons.

The MX500 offers slightly better application performance, sustained write performance, including supporting software that you can use in the US. It supports hardware encryption, and has a longer warranty for essentially the same price. The MX500 also comes in a fully metal enclosure rather than cheaper plastic.

Then there is the fact that Intel’s 660p continues to dominate in price. With prices that are lower than the NS200 at the 500GB and 1TB capacity and performance that is multiples higher than what SATA can offer, it is one of the best options out if you are looking for a cheap SSD for your next build and aren’t tied to the 2.5” form factor or SATA interface. Plus, with the 660p’s  M.2 form factor, there’s no mess of power and SATA cables to deal with.

Image Credits: Tom's Hardware

MORE: Best SSDs

MORE: How We Test HDDs And SSDs

MORE: All SSD Content

Forget Amazon Prime Day. Amazon may have a much more interesting announcement up its sleeve in the near future, if a new report holds any water. According to Bloomberg today, Amazon has several new products in the works, all surrounding its Echo brand and Alexa technology, including a new version of its smart speaker and – wait for it – an Alexa home "robot" that could potentially follow you around your home.

Speaking to sources close to the matter and briefed on Amazon's plans, Bloomberg detailed what this robot could look like. Reportedly referred to as "Vesta" internally, it's said to be about waist-high on an adult of average height and ambulates with a combination of both HD cameras and wheels. According to Bloomberg, Amazon plans to integrate Alexa in much the same way as Amazon's other Echo products, and Vesta may even carry the Echo branding as well.

The robot would presumably be able to follow you around to assist whenever needed, but there's no word on what kind of functionality it might feature. It could fall in line with other consumer robotic assistants' skill sets, such as offering the day's weather and news, allowing users to control other Alexa-enabled devices in their home or potentially acting as a base of operations for anyone looking to connect with friends and family members by way of voice or video calls.

While many are used to having speakers in their home (that may or may not be recording everything you say), it's another affair entirely to invite a more humanoid robot into your home and have it follow you around. Shoppers may not warm to this immediately if Amazon does decide to go this route, but it certainly could make for some interesting spin-off products debuting.

According to the report, the robot isn't ready for mass production just yet; however, Amazon has been pulling engineers from other established projects so that they could work on bringing the robot to fruition. It's possible that we could very well see Amazon introducing the robot in the near future.

Thermaltake (TT) is giving its Level 20 RGB gaming keyboard a Razer Green twist. In January, it announced the Level 20 with Cherry MX keys, a whole lot of RGB illumination, as well as a unique ability to control lighting via smartphone or tablet. Today, TT upped the ante with the Level 20 RGB Razer Green version. Naturally, it offers Razer Green mechanical switches. It also has a slightly lower cost of $129.99 (compared to $139.99-$149.99 for MX switches). 

The Razer Green switch has a tactile bump with a distinctive click sound and optimized actuation and reset points. That means users can expect to hear and feel every key press. This compares directly with Cherry MX Blue switches offering similar feel and sound (tactile feedback with a click) without the bump. The Razer Green switch takes 50g of force to actuate with an actuation point of 1.9mm and a total travel distance of 4mm. According to Razer, those switches make it the "best in class" when it comes to gaming and typing performance. 

Additional features on the Razer-branded switch include dual-side walls said to increase stability and help protect against spills and dust. Reliability is another important factor, with the Razer switches touting an 80 million keystroke lifespans (versus 50 million on the comparable MX Blue).

Switches aside, the Level 20 RGB Razer Green keyboard carries over all of the features found in the original. This includes a 2mm thick aluminum top plate with floating keycap design, dedicated multimedia controls, audio and USB pass-through ports, 100% anti-ghosting with full key rollover and 16.8 million color RGB lighting across the front/sides. With all of its RGBs, the keyboard is able to create 13 different lighting effects on the front/sides, along with 14 other lighting effects (sides/face) for a customizable RGB illumination experience.

You can control the RGB lighting in a couple ways. One is through the Thermaltake RGB Plus software, which is able to control other TT Sync compatible devices, such as case fans, CPU/VGA water blocks, AIO CPU coolers, power supplies, LED strips and other peripherals. You can also use Razer Synapse 3 software, which works with both TT and Razer's RGB ecosystems. In addition to the software on your PC, TT RGB Plus products also support Amazon Alexa, so you can control the lights or fan speeds by talking to an Alexa-enabled device. There are also iOS and Android apps that support voice command. If sitting back in your chair and barking out orders to your PC for lighting and fan control is your thing, look no further. 

The Level 20 RGB Razer Green gaming keyboard is available now with an MSRP of $129 through TT's network of authorized retailers and distributors. TT backs the keyboard with a two-year warranty. 

Amazon updated its Alexa app for Windows 10 today with hands-free activation. The voice assistant previously had to be manually activated on Windows 10, but with the new update, users will be able to summon the virtual helper to their PC simply by saying "Alexa" when the system is running.

This update brings Alexa on Windows 10 closer to the experience on other devices and more in line with other voice assistants. Most now respond to a specific activation phrase--or Wake Word in Amazon's parlance--to spare people from having to fumble with their devices. Alexa responds to "Alexa" by default, just like Siri listens for "Hey Siri" on Apple devices and Google Assistant waits to hear "OK Google" before offering a (virtual) helping hand.

This Alexa update will be automatically installed the next time the app is launched. From that point, saying the Wake World will "invoke Alexa regardless of whether the app is running in the foreground or background," though users can still opt for push-to-talk activation.

While requiring someone to actually press a button can help avoid problems of the service accidentally recording, it also reduces the convenience Alexa can offer, so letting people choose the option they're most comfortable with makes sense.

Amazon said the updated Alexa app for Windows 10 is available now from the Microsoft Store. The update also introduces the ability to listen to Pandora radio stations via Alexa as well as "the same features people know and love about the experience, including the ability to control smart home devices, manage calendars, play music, access tens of thousands of skills, and more."  And fear not: Alexa's expansion won't be the end of Cortana.

Microsoft CEO Satya Nadella doesn’t want to be in the business of also-rans. He’d rather see the company build on top of established platforms, which is why the Cortana voice assistant has made friends with Amazon’s Alexa and Google Assistant.

Nadella expressed this viewpoint during an “invitational media day” at Microsoft’s headquarters, according to Business Insider, earlier this week. Here’s what he reportedly told the journalists who gathered for the event:

"Would it be better off, for example, to make Cortana a valuable skill that someone who is using Alexa can call? Or should we try to compete with Alexa? We, quite frankly, decided that we would do the former. Because Cortana needs to be that skill for anyone who is a Microsoft Office 365 subscriber.”

We already knew Microsoft had decided not to compete with Alexa; the companies integrated each other’s voice assistants into their respective platforms throughout 2018. But the takeaway from Nadella’s presentation is that Microsoft won’t compete in the smart speaker market at all.

Nadella reportedly intimated that Cortana will be similarly integrated with Google Assistant sometime in the future. He said that people should be able to use Cortana on Google Assistant or Alexa, according to Business Insider, just like they can use Microsoft’s iOS and Android apps. It’s not clear if the Google Assistant mention was a sly announcement or a hypothetical.

Either way, Nadella followed with this: “We are very mindful of the categories we enter where we can do something unique. A good one is speakers. To me the challenge is, exactly what would we be able to do in that category that is going to be unique?” Because he didn’t have an answer, Microsoft opted not to bother making an also-ran smart speaker. And honestly? Good on Microsoft.

Too many companies attempt to build platforms months or years after their competitors have established their own. Sometimes this works—Apple is infamous for successfully expanding into existing product categories and pretending they’re brand new. Usually, however, this approach simply leads to needlessly fragmented markets until the also-rans finally call it quits.

Microsoft’s new approach lets it build its own ecosystems on the platforms its customers already have. That makes it easier to use products from several companies, like pairing a Windows 10 PC with an iPad and Amazon Echo speakers, without having to worry about the availability of certain services for all of those devices. Many consumers prefer this to the siloed approach most companies have taken with their services.

We heard similar promises from Microsoft when it announced the Windows 10 Fall Creators Update in 2017. The company revealed the so-called Microsoft Graph alongside it. Instead of locking down features like clipboard history sharing, document versions, and the like to specific hardware like Apple does with Continuity and Handoff, the Fall Creators Update were the company’s way of making tech better no matter what devices people use.

At least now it’s clear that Microsoft continues to see this approach as viable and that it’s going to apply it to new device categories, like smart speakers, as appropriate. Better to hitch a ride on a competitor’s back than to compete in a race you’re destined to lose, right?

Razer's Chroma lighting and its Synapse program go hand-in-hand, but soon there will be another way to control Razer's RGB lights: Amazon Alexa. At CES in Las Vegas, the company announced that the voice assistant will be able to control gaming hardware with Razer Synapse 3. The feature will come to Razer Synpase 3 in Q2 in the U.S. and Canada (later in 2019 elsewhere), and those who connect to Alexa will see a module to enable it.

You don't need an Echo speaker, either. You'll be able to use Razer's headsets and microphones (Razer told me others might work, but its devices are "recommended") to control your PC's lighting, change profiles and adjust settings. And besides adjusting lighting, Alexa can still do most of the things hardware speakers do, like check the weather and answer questions.

At one point during my demo, a Razer rep asked Alexa to have Synapse throw a "Chroma party," in which music started blasting and the lights started dancing around.

Additionally, Razer showed off the growth of its Chroma Connected Devices program, which was announced last June, and how devices from 15 partners, including Thermaltake, MSI, G.Skill and AsRock can connect RGB software to Synapse to control all of your lights through Razer's software.

Vuzix Blade, AR smart glasses with the look of Oakley sunglasses and the power of Amazon Alexa and Google Assistant, are finally hitting the consumer market later this month for $999 without prescription and "under $1,200" with prescription. Armed with a quad-core ARM processor, an AR display in the right lens, a touchpad, voice control, and its own iOS and Android companion app and app store, the Blade gives you a lot of ways to control it. We've tried them out for ourselves, and while they’re still a ways off from turning users into fashionistas, its software makes the glasses stand out for all the right reasons. 

The Look

It’s hard for me to get on board with the Vuzix Blade’s thick, black plastic styling. At best, they look like an extra chunky version of traditional men’s Oakley sunglasses—which I’m not a fan of to begin with—without the tint. Sure, they’re a more socially acceptable upgrade compared to other AR/MR glasses like Microsoft Hololens or the now enterprise-only Google Glass, but they still don’t look like something I’d voluntarily wear in order to look good. And they’re not nearly as cute as North’s Focals smart glasses.

The glasses’ temples are quite wide and gently scraped across my hair and head as I pulled them on and off. While I can feel the thick pieces resting at the back of my head, they didn’t bother me. The glasses were even more comfortable than wearing a headband.

I wore the Vuzix Blade on and off for about 45 minutes. It didn’t weigh my head down or feel too heavy. But it did irritate the bejeezus out of the bridge of my nose. The nose pads are thick rubber and gave me a feeling that I can only compare to what you feel when someone points a finger between your brows but doesn’t actually touch you (if you’ve never done this before, try it. It’s exactly how my nose felt while wearing the Blade). As a result, I felt like taking the glasses off every couple of minutes and rubbing the bridge of my nose back to normalcy.

Vuzix said the end goal is for the Blade to be as sleek and functional as the glasses used in the Kingsman movies. At this iteration, I have to say that Eggsy looks better in his version of AR glasses than I do in the Blade. But in the picture below you can see Vuzix’s vision for subsequent models, the Blade 2 and Blade 3. 

How It Works

The Vuzix Blade has to be paired with your smartphone via Bluetooth to work. It always has to know your location because the device has to know it’s physically next to you in order to function, I was told. After the Blade is connected, and you download Vuzix’s companion iOS or Android app, the smart glasses can display incoming notifications from any of your smartphone apps. The companion app will also give users access to Vuzix’s app store, the only place where you can download apps, like Amazon Alexa for the Blade.

I’m sorry; I can’t not say it: Navigating the Blade’s menu worked just like Saiyan scouters in Dragon Ball Z. With my finger, I touched a nondescript plastic touchpad on the glasses’ right arm. To scroll through the menu, a row of icons on the bottom of the display, I swiped on the touchpad. A two-finger hold brought me to the Home menu, featuring the date, time and any notifications. After a while of continuous use, due to my historically limited upper-body strength, my arm got a little tired from working the glasses.

Using display Cobra engines and a small projector on the temple, the Blade projected crisp, colorful text and images in front of my eyes. The default screen showed “Home” in big white letters, while a neon green box held the date and time.The bottom of the display featured a row of colorful icons representing apps.

Already-Impressive Array of Apps

Vuzix claims to have 400 developers a month registering to develop Blade apps. In my demo, I got to try out the Blade’s texting capabilities, gaming and betas for Amazon Alexa and a top video streaming service. It gave me a taste of what the Vuzix Blade app store will have upon the Blade’s consumer launch. It’s clear Vuzix was waiting for the right amount of apps, both in quality and quantity, before releasing the Blade to the general public.

When I tried the Blade, they were still developing the iOS texting app, so I used an Android phone. When I received a text, a pop-up dropped down from the top of the display, with a preview of the message written in white, including the sender and emojis. There’s no messages app, everything goes through your smartphone so you use your phone’s SMS app, Google Hangouts, WhatsApp or another messaging client to chat.

I tried out a teleprompter and karaoke demo where words streamed in front of my eyes like the story appears in a Star Wars opening crawl. While this appealed to me as a newfound karaoke fan (ever since my six-year-old niece did a of “Ice Ice Baby” this Christmas), there are greater possibilities for this technology. Imagine doing a public speech with your words right in front of your eyes, or this being used as a translator, instantly decoding foreign languages.

When I demoed the beta video streaming app, the company was still in the process of making the UI more seamless. But the bright colors of Rainy With a Chance of Meatballs were impressive on such a tiny display from a relatively light device. However, the image did fade out a little if I was looking at something black. The final app should be ready by the end of January.

The Blade’s Alexa compatibility has gotten it a lot of attention. But with North’s Focals also boasting the voice assistant, Blade will need to work intuitively with Alexa to stand out. When I tried out an Alexa Beta demo, I had to tap to speak and then tap when I was done speaking. It worked on the first try multiple times, but all the tapping isn’t necessarily carefree. However, Vuzix told me they are working on making Alexa voice-enabled through the Blade. And the Blade will also feature Google Assistant; although, I didn't get to try that out. 

Final Thoughts

Vuzix Blade is a bit off putting to the more style-conscious due to its chunky look, but it was surprisingly lightweight on my face. While it did irritate my nose, perhaps I can get use to it, especially if Vuzix delivers enough functionality and fun through its app store. We’re also interested in testing out that 8MP camera, battery life and long-term application. We’ll be reviewing all this in the Blade in detail later this month, so stay tuned.

Amazon's digital assistant, Alexa, is now available as an app for Windows 10 in the Microsoft Store. It's a bit of a blow to Cortana, Microsoft's own assistant that simply doesn't have the same recognition.

The app, which you can download here, requires an Amazon account to sign in. And, surprisingly, it uses Microsoft's Fluent Design elements, like acrylic panels, and also works with dark mode. Currently, it's only available in the U.S., UK and Germany. It will be available in more countries in 2019.

Alexa was preinstalled on some devices by the likes of Lenovo, Acer and HP, but now you can put it on any Windows 10 device. While it can do anything your smart speaker can, you won't be able to change the wake word. You can use Alexa to shop or control your smart home, but it has no specific control over PCs.

In the case of my work-issued Dell XPS 15, I found the laptop doesn't support wake on voice at all. I had to press a button to talk to Alexa.

"The Alexa for PC app available in the Windows store does not feature hands-free support," an Amazon representatiave told Tom's Hardware.

Earlier this year, we saw the preview of a budding romance between Cortana and Alexa, in which you can ask one to talk to the other to fill in some gaps. However, that had limited integration, and we don't know when that will be final.

Earlier this week, Cortana head Javier Soltero announced he is leaving Microsoft. This could mean a change of pace for Cortana, which has lagged behind the likes of Alexa and Google's Assistant.

It's been a long time coming, but Microsoft's Cortana and Amazon's Alexa are finally working together. The feature is rolling out now.

The companies announced last August that Cortana and Alexa would start communicating with each other. Microsoft set up a web page where interested developers could receive notifications about the integration at the Build conference in May, and now, finally a preview of the setup is available to all. Just say "Hey Cortana, open Alexa" to a compatible Windows 10 PC or "Alexa, open Cortana" on an Echo device to get things going.

Apparently the idea here is to have Cortana handle everything related to productivity, like a digital secretary, while Alexa manages everything on the personal side of things. That makes sense: Microsoft can integrate Cortana with the Office suite and other apps easier than Amazon can, but developers have already created thousands more "skills" for Alexa than for Cortana. It's all about maintaining that work-life balance.

A Microsoft spokesperson told ZDNet that this integration is currently limited. Cortana will tap Alexa for shopping on Amazon, for example, while Alexa users will "have access to Cortana's knowledge and helpful productivity features such as calendar management, day at a glance, and rich email integration." Many of Alexa's third-party skills will be available to Cortana users, but not all, though that's supposed to change in the future.

That isn't exactly the most seamless of integrations--it adds one more step between activating the virtual assistant and having it do what you want--but it will make Cortana and Alexa available on more devices. (Not that Amazon's waited on Microsoft to expand to Windows 10 laptops, television sets, and other devices.) It's easier for the virtual assistants to reach as many people as possible by partnering up than by competing.

Amazon said in an August 2017 press release that the two would complement, not replace eaceh other.

“There are going to be multiple successful intelligent agents, each with access to different sets of data and with different specialized skill areas. Together, their strengths will complement each other and provide customers with a richer and even more helpful experience,” said Jeff Bezos, founder and CEO of Amazon. “It’s great for Echo owners to get easy access to Cortana.”

If you've been waiting to use Alexa when only Cortana is available, or vice-versa, this preview might be exactly what you're looking for. But we suspect that many people have yet to make a single virtual assistant part of their lives; using one to invoke another seems like a bit of a stretch.

Many people have invited Amazon's virtual assistant, Alexa, into their homes. The service has expanded beyond Echo speakers to laptops, television sets and hotel rooms that promise to offer your own personal concierge. Yet a group of hackers from Def Con 26 this week have shown once again that setting up a bunch of internet-connected microphones in both public and private spaces can have serious privacy implications.

That's because the Echo speakers that house Alexa have been hacked. Again. The hack demonstrated at Def Con 26 used multiple vulnerabilities to compromise an Echo speaker, use it to affect other speakers on the same network and then have Alexa eavesdrop without any warning.

The good news is that Amazon has already released a patch to prevent this particular hack from working. The company has been quick to resolve security issues found in Echo speakers, especially when compared to other Internet of Things (IoT) product makers, many of which simply allow their devices to be riddled with vulnerabilities. Amazon's swift response to hacks like this one is a welcome change for this sector.

Yet these hacks continue to show that it's possible to use Alexa-equipped microphones for mischief. It doesn't even have to be intentional. Remember the Portland family that discovered their Echo sent parts of their conversations to random contacts? That didn't result from someone hacking the speaker; it happened after a series of unfortunate events led to Alexa doing what it thought it was being told.

That's not all the hackers demonstrated at Def Con 26. Their notes about the presentation explain:

"In this talk, we will present how to use multiple vulnerabilities to ... remote attack some of the most popular smart speakers. Our final attack effects include silent listening, control speaker speaking content and other demonstrations. And we're also going to talk about how to extract firmware from BGA packages Flash chips such as EMMC, EMCP, NAND Flash, etc. In addition, it contains how to turn on debug interfaces and get root privileges by modifying firmware content and Re-soldering Flash chips, which can be of great help for subsequent vulnerability analysis and debugging. Finally, we will play several demo videos to demonstrate how we can remotely access some Smart Speaker Root permissions and use smart speakers for eavesdropping and playing voice."

We expect hacks like this to continue surfacing. As more companies introduce smart speakers (even Facebook is reportedly planning to make one), it's likely more people will buy them. Just like anything else with an internet connection, if these products become popular enough, a lot of effort will be spent finding ways to bypass their security. Here's to hoping none of those microphones hear anything interesting.

Everyone wants a personal concierge. You can travel without one, sure, but then you're stuck checking the weather or looking for room service menus like some kind of animal. Of course, most people can't afford to stay in places where a top-of-the-line concierge service is provided. With the new Alexa for Hospitality product, Amazon is hoping that a technological stand-in will suffice, and it's already found partners who think it will.

The company announced that its voice assistant would move in to select rooms at the Marriott Hotels, Westin Hotels & Resorts, St. Regis Hotels & Resorts, Aloft Hotels, and Autograph Collection Hotels this summer thanks to a partnership with Marriott International. So if you're planning to go somewhere later this year and you're just dying to make sure Alexa will be right there with you, well, now your hopes will be realized.

Amazon said Alexa for Hospitality would run on Echo speakers in each hotel room. The service will do pretty much what you expect. You can ask for information regarding the weather, hotel amenities, or travel services; call for room service, housekeeping, or the real-live concierge; and control internet-connected devices just like you would at home. It's the Alexa you already know with hotel-specific information programmed in.

If you do already use Alexa, sometime in the future you'll be able to sign in to Alexa for hospitality with your existing Amazon account to play music or listen to audiobooks, and Amazon said you'll automatically be signed out when you check out of the hotel. But it's not clear exactly how that sign-out process will function, and if it doesn't work perfectly, you're stuck having your account connected to a public device. That's not great.

There are some other things to worry about regarding Alexa for Hospitality. Despite the voice assistant's increasing popularity, it's become increasingly evident that not all the kinks have been worked out. The most notable issue arose when Alexa mistakenly recorded parts of a family's conversation and sent it to a random contact. Not only did it misinterpret a string of commands needed to perform those functions, but it also wasn't loud enough in its requests for confirmation for the family to realize that something had gone wrong. It was the perfect storm.

That's bad enough when it happens in your own home. Hotel rooms are smaller, which means Alexa could hear more parts of a conversation, and people often have "private time" in their hotel rooms that they don't want to share with anyone else. (Of course, being in a smaller space could help even the hard-hearing realize that Alexa's asking for permission to record and share the noises they're making with the outside world.)

You also have to wonder how the hotels plan to ensure their Echo speakers' physical security as well. Countless people have access to those devices--any security vulnerabilities found in them would be relatively easy to exploit from the comfort of a hotel room. Researchers have already demonstrated exploits in Echo products and, even if they've already been fixed, it makes it clear that these products aren't infallible.

All of which brings us right to the common question: are you willing to risk your privacy in exchange for convenience? Alexa for Hospitality doesn't offer any information you can't find elsewhere; hotels are chock full of resources meant to make planning your trip easier. Amazon's voice assistant merely lets you access that info by talking to a robot instead of leafing through pamphlets or conversing with another human.

Amazon's Alexa is coming for you. The voice assistant debuted with Amazon's Echo products, expanded to Acer laptops last week, and has now reached LG's smart TVs. LG said that Alexa will be supported with this year's OLED and Super UHD television sets, and that the voice assistant will do pretty much what you'd expect, whether it's controlling video playback, flipping through channels, or searching for specific content.

A myriad of other products, from internet-connected lights to smart locks, also support Alexa. Yet along with those expansions come new privacy concerns, as shown when a Portland family discovered that Alexa had quietly sent part of their conversation to a random contact.

Privacy-minded people have long feared that voice assistants like Alexa were little more than spies in disguise. Although microphones have been part and parcel to our digital lives with smartphones, laptops, and other devices for years, many of those products weren't always listening and waiting to be activated. You had to answer a phone call, for example, or use programs that made it obvious when your microphone was actually being used.

Alexa is supposed to do the same, of course. Compatible devices are set up in a way that's supposed to ensure that listening for the activation phrase doesn't actually record every sound made anywhere near the microphone. Think of it kind of like zoning out during a boring lecture: you're ignoring whatever that lecturer is talking about until you hear your name, and then you start to actively listen and pay attention to what they say.

But it doesn't always work that way. KIRO 7 reported last week that a Portland family discovered that, unbeknownst to them, Alexa sent parts of their conversation to a contact. Danielle (whose last name was kept private) said one of her husband's employees called from Seattle to say their Alexa devices had been hacked and were sending audio snippets to that employee. That isn't exactly what happened, though, as Amazon told KIRO 7:

“Echo woke up due to a word in background conversation sounding like 'Alexa.' Then, the subsequent conversation was heard as a “send message” request. At which point, Alexa said out loud 'To whom?' At which point, the background conversation was interpreted as a name in the customer’s contact list. Alexa then asked out loud, '[contact name], right?' Alexa then interpreted background conversation as 'right.' As unlikely as this string of events is, we are evaluating options to make this case even less likely.”

We don't know how often scenarios like this happen. It's possible that Alexa has never misinterpreted as much "background conversation" as it did in this instance, but as the voice assistant reaches more and more devices, it becomes a nigh-ubiquitous network just waiting to hear "Alexa." Your laptop, TV, wireless speaker, tablet, and other gadgets and gizmos are increasingly listening to everything you and the people around you say.

Most of the time that listening is harmless. But freak accidents like the one described above do happen, and eventually they'll involve more than harmless conversation about hardwood floors being sent to an employee kind enough to warn their boss about the intrusion of privacy. Alexa-compatible devices already call bedrooms, living rooms, and kitchens home; expanding to portable devices like laptops will let them wander, too.

Those fears could be compounded by intentional eavesdropping instead of accidental broadcasting. A security company called Checkmarx revealed in April that a flaw in Amazon's Echo products could let Alexa spy on its owners at all times without their knowledge. Attackers could use malicious apps (or "skills" in Amazon's parlance) to keep an Alexa recording session going long after its owner thought it had stopped listening.

Many Alexa users are probably happy to learn that the assistant is expanding to Acer's laptops and LG's smart television sets. But the news from Portland that was sandwiched in between those announcements should give people pause. At the very least, it's worth noting that you're inviting more "hot" microphones into your lives every time Amazon's assistant reaches a new device. Is the convenience worth the risk? Well, that's up to you.

Move over Siri, Google and Cortana. Here comes Alexa. Amazon's take on the digital assistant is making the jump to PC, starting with Acer.

In fact, the laptop manufacturer will be the first to offer Alexa, as the assistant will be preinstalled on several of its 2-in-1 laptops including the Spin 3 and Spin 5 starting on May 23 and May 25, respectively.

Nowadays, it's all about the internet of things, and more specifically digital assistants like Amazon Alexa. To date, you can find Alexa in tablets like the Fire HD 8, soundbars (Polk Command Bar), headphones (OnVocal OV earphones) and of course, proprietary Amazon products like the Echo, Dot, Show and Spot. It was only a matter of time before the technology made the leap to laptops.

MORE: Best Acer Laptops

In addition to the Spin 3 and Spin 5 notebooks, Acer will also add Alexa to systems like the gamer-centric Nitro Spin 5 (update available starting in June). The company also plans to bring the technology to several laptops in its popular Swift, Aspire and Switch lines of notebooks as well as all-in-one PCs in coming weeks. If you already own one of the designated systems, you can expect an update via Acer's Care Center software sooner than later.

Consumers can summon Alexa using Intel's integrated digital sound processor, Intel Smart Sound Technology, which is specifically designed to handle voice and speech interactions. Some of Acer's laptops, such as the Spin 5, have four integrated microphones and far-field technology which will allow consumers to summon Alexa over fairly long distances, even in slightly loud spaces. Other Acer laptops will also utilize far-field tech, but will only have two mics to work with.

MORE: Best Alexa Skills

Similar to other digital assistants, Alexa will let consumers check the weather, update their schedules and play music among other commands. Alexa can also be used to control compatible smart home products, allowing you to adjust the temperature, turn on the lights or activate a smart security system from just about anywhere. If you don't own an Acer laptop, fret not -- the assistant is slated to come to a range of Windows 10 laptops throughout the year.

Want to learn what else Alexa can do? Be sure to check out our comprehensive guide to Alexa on our sister site, Tom's Guide. 

This article originally appeared on our sister site, Laptop Mag.

Input Club is now taking orders for its newly-designed Hako switches. Invented by Input Club members Jacob Alexander and Brandon Muzzin, the Hako switches are a variation on Kaihua’s Box switches and feature a vented, self-cleaning mechanism and a more stable slider that results in reduced wobble, purportedly increasing lifespan by 25%.  

Like the Halo switches designed by Input Club and produced by Massdrop, there are two version of these new switches, a “True” and a “Clear.” Both are medium-weight tactile switches designed to create a smooth actuation curve similar to Topre’s 55g switch.

The Hako Clear features a standard spring with the tactile event occurring at ~0.66mm at an actuation force of ~55g, with a peak tactile force of ~63g. It uses a plastic half-cylinder and metal leaf design, giving it a defined tactile bump much like the Cherry Clear switches they emulate.

Though similar in construction to the Hako Clear, the Hako True switches feature a shortened spring with the tactile event occurring at ~0.54mm, and a slightly greater actuation force of ~60g. However, a lower tactile force of ~58g makes for a more subtle tactile bump than its sibling. The True switches also feature a lighter preload on first touch.

Both switches feature ~3.6mm of travel with actuation at ~1.95mm. There’s a tension bias at the end of the travel, peaking at ~94g in the Trues and ~79g in the Clears, that’s intended to discourage you from bottoming out the slider when you type. The switches are rated for 80 million presses and are compatible with both through-hole and SMD LEDs. Both have clear housings.

Order buy-in is $54, which gets you 120 switches of either type. The group buy runs until November 28, and shipping is set for after the holiday season, on February 18 of next year.

Table source: (True, Clear)

Update, 5/18/17, 7:13am PT: In the original version of this article, we did not have a force curve graph for the Halo Clear switches. We've now acquired it, included it below, and added to/adjusted the copy accordingly.

Original article, 5/17/17, 5pm PT:

Disclaimer: Members of the Input Club have written for Tom’s Hardware.

Update, Mechanical keyboard enthusiasts are always keen on new switches, and the Input Club has served up two more. Released along with its K-Type keyboard, the two new switches were designed by Jacob Alexander of the Input Club and are manufactured by Kaihua Electronics, which makes Kailh switches.

Halo True

The Halo True is intended to improve upon the Cherry MX-style slider. On its Halo True page, the Input Club describes it thusly:

The true innovation invented with the Halo is the recreation of an entirely smooth curve, combined with a lack of pre-load or tension on the spring at rest. With normal switches, there is a small amount of friction present when you first press down. With the Halo, this is largely absent and you are able to enjoy a smooth press from beginning to end.

The Input Club basically wanted to build a switch that’s as smooth as a Topre switch but eschews Topre’s conical spring for a more Cherry-ish design. You can see the force curve of the Halo True here:

Just as we draw some conclusions from our own switch testing, you can see quite a bit in the Input Club’s Halo True force curve. Although this is a tactile switch, the overall curve looks relatively smooth. Note, for instance, that the peak force is about 60gf, but the actuation point (which is well after the peak tactile force) is just 52gf. That’s a delta of just 8gf.

Further, just after the actuation point, the force bumps up to about 62-64gf. Thus, if you were to draw a line from the peak of the tactile bump to the beginning of that force increase, it would be a nice, gentle slope.

Also note that this is a fairly heavy switch; 60gf on the peak of the tactile and 52gf at actuation is significant enough, but the Halo True requires a whopping 100gf to bottom out. This, compared to Logitech’s Romer-G switches, which we discovered had a peak force of ~48gf, actuation force of ~43gf, and bottom-out force of only ~60gf.

It's also important to bear in mind, though, that the Input Club designed this switch so that you don't need to bottom out. The design, in fact, specifically encourages you not to bottom out the switch, so you have a nice, bouncy keyfeel when typing.

Halo Clear

The Halo Clear switches are designed to have the same spring weight as the Cherry MX Clears with the smoothness (the Input Club calls it a “velvety sensation”) of the Halo slider. In other words, the two Halo switches have the same slider but different springs.

The Halo Clear, you may have surmised, has a clear switch housing for RGB lighting. The Input Club stated, “This switch was invented for the K-Type, to fill the void made a lack of RGB capable switches similar the Cherry MX Clear.” 

The Halo Clear has much lower spring force and bottom-out force than the Halo True, so you can expect to bottom out these switches more. It's also interesting that although the switch specs indicate a full travel of 4mm, the force curve chart above shows that the travel ends a little closer to ~3.7-3.8mm.

This switch will have a much more tactile keyfeel than the Halo True. The peak tactile force is at 65gf, and it dips all the way down to ~47-48gf just before actuation. Then, right at the actuation point, the force bumps up again to about 60gf.

Both switches are currently available only on the K-Type, which you can order from MassDrop (for the next thirteen days or so). However, eventually you may be available to acquire batches of the switches so you can build your own keyboard with them.

Websites are only as safe as their operators allow them to be. Researchers find vulnerabilities, and organizations release patches for them all the time. But it's up to site operators to install those patches to make sure their sites don't endanger visitors and, potentially, their personal information. Many don't, as Northeastern University discovered when it found that 37% of sites use outdated JavaScript libraries with at least one known vulnerability.

The Northeastern University researchers scanned 75,000 of Alexa top websites and 75,000 more websites randomly taken from "a snapshot of the .com zone in May 2016." They found that more than one-third of Alexa's top websites use one JavaScript library with a known vulnerability. Another 10% use two or more libraries with known security problems. Roughly 97% of the sites wouldn't be able to fix all their vulnerabilities with a few quick patches.

This means that many website operators who haven't bothered to regularly patch their JavaScript libraries will now have to install whole new libraries to address all their security vulnerabilities. That seems unlikely at best--many of these websites use outdated libraries for a reason, and it's probably not because their owners are infatuated with old tech. It's most likely because the sites have been abandoned or simply aren't made with security in mind.

Not that JavaScript library creators and maintainers have made things easier for website operators. As the researchers said in their paper, titled Thou Shalt Not Depend on Me: Analysing the Use of Outdated JavaScript Libraries on the Web:

Perhaps our most sobering finding is practical evidence that the JavaScript library ecosystem is complex, unorganised, and quite 'ad hoc' with respect to security. There are no reliable vulnerability databases, no security mailing lists maintained by library vendors, few or no details on security issues in release notes, and often, it is difficult to determine which versions of a library are affected by a specific reported vulnerability.

Those aren't small problems. The researchers said that 87.7% of Alexa websites and 46.5% of the randomly selected websites use "at least one well-known JavaScript library" such as jQuery, Handlebars, and Angular, among others. These open source libraries are used by millions of sites that weren't analyzed by Northeastern University; it's hard to imagine that you've never visited a site that uses at least one of them.

Vulnerabilities in these and other JavaScript libraries can cause all sorts of headaches. Website operators might worry about someone exploiting a vulnerability to inject code on their site; you might worry more about someone abusing a security flaw to gather your personal information. The researchers explained:

Attackers can use these capabilities to steal data from a user’s browsing session, initiate transactions on the user’s behalf, or place fake content on a website. Therefore, JavaScript libraries must not introduce any attack vectors into the websites where they are used.

None of that sounds like fun. But the researchers also found that many popular JavaScript libraries don't share any security problems with devs:

During this entire work, we did not encounter a single popular library that had a dedicated mailing list for security announcements (in fact, most libraries we investigated did not have a mailing list for announcements at all). Furthermore, only a few JavaScript library developers provide a dedicated email address where users can submit vulnerability reports. When the release notes of libraries mention at all that a vulnerability has been fixed, they often do not provide any details about the affected code, or which prior versions are vulnerable.

The end result: nigh ubiquitous JavaScript libraries that don't help website operators keep you safe. The websites themselves, their users, and personal information about an untold number of people could all be vulnerable to attack simply because developers didn't know they were using outdated JavaScript libraries or weren't sure if using an old version of the tech would undermine their website's security.

Introduction

What, exactly, does a router do? And how does it do it? Although the answer to the first question can be relatively simple—a router routes data packets along networks—all of the complexity lies in "how."

Routers come in many shapes and sizes: a room full of rack-mounted, dedicated systems for enterprise clouds; off-the-shelf commercial boxes from manufacturers like Linksys or D-Link; and DIY solutions the size of a deck of cards, built on hobbyist platforms like Raspberry Pi.

Leaving aside the overall meta architecture of packet exchange, the core hardware of a modern commercial router (aimed at the small business and home networking markets) distinguishes it from enterprise or application-specific solutions. Specifically, such a router emphasizes convenience. Single devices act as a DSL modem, router, wireless access point, media server and connection to the smart kitchen sink. Also, the processor and connectivity are geared toward a completely different profile (high-bandwidth use for gaming and streaming, along with simultaneous connections from multiple devices, including smartphones and fridges) than enterprise-oriented platforms.

This article focuses on the hardware that runs modern consumer routers. And in today's routers, SoC (system on a chip) solutions are universal—all of the hardware we cover comes in the form of integrated SoCs. These multi-function systems comprise a variety of configurations with different capabilities, making it more difficult to dig into their respective architectures; but they simplify the router design process. There are far fewer devices to consider when a single board comes with everything built onto it.

That doesn't mean completely integrated systems are the only off-the-shelf setups worth considering, though. Even in the world of SoCs, factors like cost, power consumption and OEM requirements do create a world of chimeric SoC solutions, with multiple high-power radios or transceivers pressed in to service a higher-level processor/memory board, or an xDSL modem connected to one of the ports of a more general-purpose processor. Solutions with discrete RAM or flash memory modules and transceivers with separate radio chips do exist, but they are few and far between. And so we focus on the two classes of SoCs that are most often seen in the wild: processor/all-in-one chips and transceivers.

A Note On Software/Firmware

Commercial routers overwhelmingly favor Linux as an operating system, and often employ a customized version. A lightweight Web server is almost always installed as well for user-controlled device configuration.

Another OS, VxWorks, is used only in enterprise-class systems, but it merits a mention if only because it is used on the Mars Reconnaissance Orbiter, and I would buy a commercial router running VxWorks regardless of how unwieldy the final system becomes.

Finally, an open, Linux-based standard, OpenWrt, is being embraced by more OEMs, even though it was previously mostly used by prosumer hobbyists.

MORE: Wireless Routers 101
MORE: All Networking ContentMORE: Networking in the Forums

Router SoC Functional Components

In a router SoC, packetized data arrives through input ports, is directed using a network of connections called the Switching Fabric (which can be thought of as wires connecting each component to every other), and leaves through output ports.

The router SoC keeps, in memory, a look-up table of addresses associated with each packet, and the processor uses various rule sets to determine the best path for sending data. It also listens in to keep tabs on network traffic, updating its "available" and/or "best" path for packets based on load levels.

The SoC's on-board memory maintains routing tables, consisting of network and host addresses. To determine the best path for delivering a packet to its destination, the routing table includes all known network addresses, instructions for connecting to other networks, possible paths between routers and a measure of distance between nodes or network addresses known in the form of cost functions. This on-board memory usually takes the form of flash  or EEPROM. The size of the memory chip is less important than its speed.

By their nature, routers do not communicate directly with end devices like laptop or desktop computers, but with their network adapters. Each NIC has a network address—Ethernet switches, adapters, Wi-Fi transceivers and radios are all NICs.

Additional hardware can provide dedicated support to the main CPU. A cryptographic chip can offload encryption and hashing functions from the main processor. A dedicated load balancer can jump in to optimize different types of data streams—streaming video versus regular Web page browsing, for example. USB, SATA and other types of capabilities on routers are also added in the form of interface cards, often integrated right into the main processor board.

The architecture gets more complex when you consider that many routers also provide an xDSL, cable or cellular data modem. In this case, the router's interface card is the component that mediates data between the modem and switching fabric.

A modem modulates and demodulates digital signals into analog and vice versa, to be sent out over the "line"—cable or xDSL. However, we will not cover cellular modems in this article; those require architecture and protocol different from xDSL or cable modems.

Processors And Transceivers

The Processor

Any host processor can, theoretically, act as a networking processor, but the hardware in commercial routers is optimized to handle very specific networking tasks. These include key look-up (database look-up using a key), computation, data bit-field manipulation, queue management, pattern matching and control processing. A GPU, for example, would yield less-than-optimal results as the brains of a router (though the reverse problem, gaming on a networking processor, would run up against fundamental limits pretty quickly).

Before HD (or 4K) streaming video, multiple devices communicating over the same home network or intensive network gaming, wasting a multi-core processor on routing tasks was unheard of, even in the prosumer segment. Clearly, this has changed. Processor performance, especially on older routers, can bottleneck the bandwidth of ISPs serving gigabit-class Internet.

Beyond cores and clock rates, CPU specifications can quickly become mind-numbing. But in the networking world, they can generally be split into their instruction set architectures (ISAs): ARM and MIPS.

Of course, there are other architectures other than ARM and MIPS used in networking devices. These are usually dedicated co-processors like DSPs, cryptographic processors, media accelerators and so on.

ARM

ARM Cortex-A refers to a series of microprocessors designed by ARM Holdings PLC. The company doesn't manufacture the hardware, but instead licenses its designs. The suffix "A" stands for "applications," hinting that the A series is meant for general-purpose use. There are also "M" (for microcontroller) and "R" (for real-time) families.

The implementations most commonly found in modern networking devices are the older Cortex-A9 or Cortex-A5 chips. Both are based on the 32-bit ARMv7 architecture.

The Cortex-A9 was introduced in 2007. It features L1 instruction and data caches that can be configured independently to 16, 32 or 64KB; up to 8MB of L2 cache and clock rates as high as 2GHz. Cortex-A9 processors are included in many current SoCs, including the Apple A5 and A5X; Broadcom BCM11311; Nvidia Tegra 2, 3 and 4i; and even Sony's PlayStation Vita.

The Cortex-A5 hit the market in 2009 as a less powerful alternative to the A9 for low-end and mid-range consumer devices. It's available with between one and four cores, and includes from 4KB to 64KB of L1 instruction/data cache. It is used on many transceiver SoCs.

MIPS

The MIPS architecture (short for microprocessor without interlocked pipeline stages) was introduced in 1981 by John L. Hennessy of Stanford University. It is currently developed by MIPS Technologies, which has been part of the UK-based Imagination Technologies group since 2013. MIPS uses a reduced instruction set computer (RISC) architecture, enabling specialized chips with low power consumption that are widely used in embedded systems for routers. There are two instruction set versions currently in use: the 32-bit MIPS32 and 64-bit MIPS64. Both were introduced in 1999.

MIPS32 is used in various microarchitecture families—namely, 4K/E, 24K/E, 34K, 74K, 1004K, 1074K/f, microAptiv, interAptiv and proAptiv. The latter three are the most current, introduced in 2012.

The microAptiv, interAptiv and proAptiv microarchitectures typically come with 32x 32-bit general-purpose registers (up to 64x are allowed). The architecture allows for up to 8MB L2 cache, and current implementations operate at frequencies as high as 1.5GHz. MIPS32 is meant to be upward-compatible with MIPS64, which means its features are supposed to be a subset of what MIPS64 offers. Both utilize fixed-length commands in a three-operand format and a load/store data model, catering to high-level programming languages.

Chips based on MIP32 include Broadcom's BMIPS3000, BMIPS4000 and BMIPS 5000; BCM53001 and BCM1255; Ingenic Semiconductor XBurst 1; and Baikal Electronics P5600. Chips using the 64-bit MIP64 architecture include Broadcom's BCM1125H and BCM1255; the Cavium octa-core processors CN30xx, CN31xx, CN36xx and CN38xx; Octeon Plus: CN5xxx, Octeon II: CN6xxx, and Octeon III: CN7xxx; Ingenic Semiconductor XBurst 2, NEC VR4305 and VR4310.

Transceivers

A transceiver is a device that combines a data transmitter and receiver in a single package. Most commercial routers have transceivers with integrated radio/antenna systems, and most often, a small controller or dedicated chip on the board handles radio functions. While "transceivers" are most commonly associated with radio frequency (RF) signals, they have analogs in other transmission standards, though few (if any) of these analogs will be seen in a commercial router for home/small-business use. The Ethernet versions of transceivers are MAUs (medium attachment units), while fiber and 10GbE have their own set of acronyms (GBIC, XAUI, etc.) to describe transceiver devices.

Although the trend toward complete SoC solutions continues unabated, many new devices tend to take the transceiver plus processor card approach. Ensuring backward compatibility with older 802.11b/g/n networks—and avoiding a complete redesign with a router refresh, where new functionality can be thrown in just by adding another card—is a big factor. It should also be noted that many packages called "SoCs" omit the actual transmission-receipt function, instead providing a PCI/PCIe slot for the router manufacturer to fill with a network card.

Chipsets

Chipsets

The landscape of networking device manufacturers has changed drastically within the past five years. The big names—Ralink, Ubicom, Atheros—are gone, absorbed by bigger companies. The ground is now held by two giants: Broadcom and Qualcomm. A relative newcomer to the field, MediaTek, is holding its own through the strength of acquisitions. Smaller manufacturers like Marvell and Quantenna have also introduced SoCs used in many mid- and high-end routers, while the networking division of RealTek supplies D-Link and Huawei with chips.

In 2015, there were some new technologies that affected the giant-dominated line-up, including 5G Wi-Fi Wave 2, 4x4 and 8x8 MU-MIMO. All of the players have exciting offerings, and some of the smaller companies (Quantenna, specifically, with its 10G SoCs) had a lead on Broadcom and Qualcomm at CES 2015.

Chipset Vendors: Broadcom

Broadcom is the wireless and broadband component manufacturing industry's leader. Its undisputed strength lies in its transceivers and radios, which are even used in conjunction with competing router SoCs.

Broadcom Router SoCs

Introduced in 2013, the StrataGX BCM5862X series is part of Broadcom’s Northstar Plus family, featuring single- or dual-core ARM Cortex-A9 processors at 1.2GHz. They connect to two transceivers via PCIe, and are designed for 5G Wi-Fi in combination with capable transceivers. With two SATA 6Gb/s interfaces, a cryptographic accelerator and fast memory, the BCM5862X series is intended for storage appliances.

The BCM5301x and BCM470X lines brought 802.11ac-specific support to the Broadcom portfolio in 2012. The 5301x is the enterprise/small business version of the 470x series suitable for residential routers and gateways, and both are manufactured on a 40nm processes. The 4709/AO/CO chips succeeded the popular 4708 series; the AO model was released in 2013, followed by a base model in 2014. The newest member of the family, the BCM4709CO, came out in 2015 (and is also listed as the BCM47094).

Broadcom's latest and greatest for "affordable and mid-tier" residential/small-business routers, the BCM47189 and the BCM53573, were introduced in January 2015 at CES. They are (probably) meant to be paired with the BCM4366 5G Wi-Fi 4x4 MU-MIMO transceiver (radio) chip. Broadcom also announced its enterprise- and cloud-oriented SoCs from the same family, the 43465/43525 and the 47452.

An ADSL router SoC, the BCM6318 is designed for integration with entry-level networking devices, providing an all-in-one solution for turnkey router development.

Another integrated modem/router solution with an external wireless NIC, the BCM63268, is one of the most widely used chips for xDSL platforms. It is designed to work in conjunction with the BCM2057 radio. The 63168 provides VoIP support (multi-channel HD voice) and is designed for use with the BCM4360 5G Wi-Fi transceiver.

The BCM4706 sits right at the edge of obsolescence, but a number of popular, low-end routers from 2012 and 2013 use this chipset. Its dual PCIe interfaces are designed to pair with a BCM4331 802.11n 3x3 Wi-Fi transceiver, and it operates both in selectable and simultaneous dual-band configurations.

Broadcom Transceivers

The BCM4352, BCM4360, BCM43526 and BCM43516 were a family of Gigabit 5G chips introduced in 2012 meant for the consumer market. They differ mostly in nominal speeds; the BCM4360 is the fastest, at 1.3 Gb/s via three streams, the BCM4352 and BCM43526 offer two streams at 867 Mb/s and the BCM43516 comes in last, providing 433 Mb/s. The BCM43526 comes with USB instead of a PCIe interface and targets set-top boxes and televisions. The BCM4360 is widely used in wireless routers like Asus' RT-AC56S and RT-AC56U; D-Link's DIR-860L; Linksys' EA6200, EA6300 and EA6350; and Netgear's D6200, EX6200 and R6200, among others.

The BCM43131 is a Wi-Fi chipset from 2013 with PCIe interface and support for Wi-Fi standards 802.11b, 802.11g and 802.11n. Equipped with a BCM2057 radio chip, it is used in wireless routers from Tenda, namely the D152, W150D v6 and W311E.

An earlier incarnation of this chipset, the BCM43217 (2012) is equipped with the BCM2055 radio, and supports b/g/n. It was used in wireless routers from a number of vendors, including Asus' RT-AC56S and –U; Belkin's F9K1113 and 1118; Linksys' EA6200, EA6300, EA6350 and EA6400; and Netgear's DGN2200, R6200 and R6250.

Introduced in 2014 as a 3x3 MIMO 802.11ac chip for routers, the BCM43602 is equipped with a 320MHz single-core processor and comes with a PCIe interface. Broadcom rates its peak performance at 900 Mb/s. The chip has been used in multiple routers, including Asus' RT-AC3200; the Linksys EA9200; Netgear's D7000 and R8000; and the TP-Link Archer C3200.

Announced in January 2015 as Broadcom's then-fastest 4x4 MU-MIMO chip, the BCM4366 is a 5G radio unit meant to be used in high-end consumer devices. The BCM4366 comes with an 800MHz Cortex-A7 processor. Its Wi-Fi speed is a nominal 5.4 Gb/s (in theory, since, in practice, most consumer devices are unable to saturate it). More important is the MU-MIMO capability, allowing the BCM4366 to handle up to eight clients simultaneously. The radio spans 160MHz, which, unfortunately, also means that only two such networks fit in the 5GHz band without overlap.

Chipset Vendors: Qualcomm

Atheros absorbed Airgo Networks in 2006, making it a heavyweight in the wireless domain. Qualcomm, originally not a player in the commercial networking field, announced a takeover of Atheros in 2011, and Atheros became a subsidiary named Qualcomm Atheros. Qualcomm Atheros went on to acquire Ubicom for its SoC IP in 2012, and Wilocity in 2014 for its 802.11ad expertise. Another interesting Qualcomm Atheros acquisition from September 2011—Bigfoot Networks, a manufacturer of networking solutions for gaming applications—struck out on its own as Rivet Networks. Its SoCs and NICs are marketed under Killer Networking's name, and show up in high-performance gaming motherboards.

Qualcomm Router SoCs

The IPQ40X8/X9 is Qualcomm's latest Wave 2 MU-MIMO SoC. Because it was introduced in October 2015, information on this chipset is still scarce. But we do have some of its specifications.

The A/IPQ806X chipset family is designed to enable "smart-home" platforms, and it borrows from the MSM8974's mobile pedigree to do it. The 2012 APQ8064 used the same Snapdragon S4 processor as the MSM8974 and had a 3G/4G modem, whereas the IPQ8064/62 looks more like a traditional router platform but with a plethora of slots and ports, as well as SDIO support.

We wouldn't normally include the QC401X and QCA4531 chipsets, since they're targeted at low-power devices for IoT networks, but their newness merits a mention. Qualcomm’s newest WiSoCs, the low-end RTOS-driven QCA401X family and the QCA4531 SoC that runs OpenWrt Linux, support the AllJoyn IoT standard running off a 650MHz, MIPS 24Kc-based processor.

A very popular chipset family, the QCA95xx was introduced in 2013 and was refreshed in 2014. It's found in routers from pretty much every vendor.

Qualcomm Transceivers

The QCA9880 is a 3x3 dual-band radio chip introduced in 2013. It's meant to be paired with the QCA9558 SoC, providing up to 1.7 Gb/s. It is used in various routers, including Cisco's DPC3941; D-Link's DIR-859, DIR-862L and DIR-863; Linksys' E8350; Netgear's C6300 and R7500; and TP-Link Archer's C5, C7, D7, TGR1900 and TL-WDR7500.

The QCA9882 is a 2x2 dual-band radio chip introduced in 2013; its the QCA9880's "little brother," despite its higher model number, and it complements the QCA9880 for home networking. The QCA9890 and QCA9892 are their counterparts for enterprise solutions. The QCA9882 is rated for up to 1.3 Gb/s and is used in various routers, including Asus' RT-AC55U and RT-AC55UHP; D-Link's DAP-2660 and DGL-5500; and Netgear's D6200, JR6100, R6000 and R6100.

In the same family, the QCA9890 and the QCA9892 were also introduced in 2013 as 2x2 and 3x3 dual-band radio chips. They both provide up to 1.3 Gb/s utilizing the 802.11ac standard. The difference between them lies in their number of streams. QCA9890 is the "bigger" of the two, complementing chips for enterprise solutions featuring three streams. The QCA9890 only offers two streams. The QCA9890 has been used in AirTight Networks C-75 and -E, as well as Gateworks Ventana GW3056.

The QCA9860 and the QCA9862 are stand-alone combo chips that complement the above families of radio chips introduced in 2013. The bigger of the couple is the QCA9860, offering three streams, whereas the smaller one, the QCA9862, only offers two. Both reach up to 1.3 Gb/s. Unlike the QCA9880/82 and QCA9890/92, which are meant to be paired with an SoC solution, the QCA9860 and QCA9862 are stand-alone SoCs.

Chipset Vendors: MediaTek

Taiwanese chipset manufacturer MediaTek started out in the optical drive and home entertainment segments, then moved to dominate the smartphone and mobile chipset markets. In 2011, the company bought Wi-Fi chipset manufacturer Ralink, whose chips could be found in every router vendor's devices, marking its entry into the networking chipset space. Ralink itself had previously purchased its major competitor TrendChip in 2010, acquiring ADSL SoC expertise.

MediaTek Router SoCs

MediaTek's integrated SoC offering, the MT7623A/N, was announced in Q2 2015, with optimizations for audio/video streaming. With a storage accelerator and the OpenWrt standard, this chip has the flexibility to enable very capable NAS setups as well.

Intended for IoT gateways and media routers, the MT7683 was announced in Q3 2015, and it differs from the MT7623A/N systems in some key areas—noticeably, the introduction of a Mali 450 GPU. This allows the 7683 to display the status of connected IoT devices on a monitor or TV. IoT control is provided by the MT7687 SoC, MediaTek's first ARM Cortex-M4-based IoT Wi-Fi solution.

The MT7683/23 chips support a number of content streams over cable, Bluetooth and BLE for wearable devices. NFC is enabled for quick setup. Wi-Fi is delivered via the powerful 802.11ac Wave 2 MTC7615 transceiver, announced in Q1 2015.

A power-efficient IoT SoC, the MT7687, was announced in Q2 2015. With a maximum power output of 21 dBm, this chip works as a stand-alone IoT gateway or with the MT7683 as a powerful smart-home solution.

MediaTek's most popular offerings are MIPS-based SoCs. The MT7621 A/N/S powers everything from mid-tier routers to access points. Another comprehensive low- to mid-range SoC, the MT7620, is also used in a variety of networking applications, and is extremely popular across all market segments. The MT7628 family is an update to the popular 7620.

An integrated xDSL (VDSL2/ADSL2+ IAD) and router solution, the MT751x series, is designed for a flexible networking system design, also containing a little bit of everything. Interestingly, these chips adopt a twin-CPU solution consisting of a 32-bit MIPS CPU and an xDSL Discrete Multi-Tone (DMT) engine.

Although Ralink was absorbed by MediaTek, its last few chips were showing up in routers as late as 2013. And the sheer number of devices powered by Ralink silicon means you can't quite forget about the company's SoCs. The 6855 was the last Ralink chip to show up for FCC approval in 2013. Both the 6856 and the 6855 were powered by the dual-core MIPS 34KEc 700 processor. The RT63XXX family of xDSL router SoCs were still being used for new devices as late as 2014 by a loyal TP-Link (in one case, married to a MediaTek transceiver; TP-Link's TD-W8951ND v6 was powered by Ralink's RT63365E and MediaTek's MT7601E).

Smaller Chipset Manufactures

Marvell Technology Group creates SoCs for networking devices, often pairing them with Broadcom transceivers. Chips in Marvell's Armada 38x family are equipped with ARMv7 Cortex-A9 dual-core processors (with the 88F6810 chip, Armada 380, being an exception). They feature GbE, DDR3/3L/4, PCIe 2.0 links and a host of other features that make Marvell competitive. We're not seeing many devices with the company's hardware, though. Armada XP (MV78XXX) chips, with up to a quad-core ARM v7 PJ4 processors, seem to share the same fate.

A new addition to Marvell's Avastar SoC family was announced in Q1 2015, targeted at enterprise APs, hotspots and residential multi-stream (video or gaming) applications. An earlier iteration, the 88W8864, supported up to 1.3 Gb/s and 4x4 MIMO. It was used in the Linksys WRT1200AC and WRT1900AC wireless routers.

We're calling it a "smaller" manufacturer, but in reality, Realtek is one of the largest chip makers in the world. But the company's networking segment peaked with its 10/100 Ethernet controllers. Its wireless offering haven't enjoyed the same level of market penetration. So it's not the most popular chip on the block, but it does deserve a mention: the 2013 ADSL2+ modem/router from RealTek, based on its Lextra LX processor (a 32-bit implementation of the MIPS architecture) held its own for a while. It was used in D-Link's DSL-2740E, Huawei's WS319 and other routers.

Quantenna specializes in wireless SoC transceivers. Its R&D focused on high-end 802.11ac and 802.11n devices. Competing neck and neck with the giants of the industry, Quantenna introduced a number of new devices in 2015, all on the cutting edge. It was first to launch a 4x4 MU-MIMO 802.11ac chipset, and has demonstrated a 10G system. The QSR2000 Wave 2 is a transceiver designed for high-speed Wi-Fi routers. It is marketed as an integrated chipset for 802.11an/ac or 802.11b/g/n Wave 2 applications, dual-band switchable, with 4x4 MU-MIMO four spatial streams. It has 80MHz channels (for the 5GHz band), PCIe 2.0 connectivity and a peak PHY of 1700 Mb/s.

The QSR10G family of chipsets supports 10Gb speeds. There are four variants, ranging from the "U"—a top-tier 12-stream, dual-band device with a peak PHY rate of 10 Gb/s—down to the "5" variant, with eight-stream 5GHz single-band operation with a peak PHY rate of 8.6 Gb/s.

The Future Of Router SoCs

Just as in-home HD video and game streaming drove the innovations behind MU-MIMO and Wave 2 wireless, IoT and smart-home initiatives are poised to drive capabilities in the next generation of routers. Expect to see more and more devices capable of handling low-power, low-data rate, always-on clients in addition to existing high-end capabilities. Qualcomm is already making forays into this area, and Broadcom will not be far behind. Intel is also collaborating with cellular modem manufacturers and has a grand IoT vision, so expect to see new players in the field.

Another series of innovations will target mobile routers—those with integrated cellular modems. With the number of travelers carrying two or more computing devices (laptops, smartphones, tablets, smartwatches), the demand for small, integrated wireless router/cellular modem combination devices is expected to rise. MediaTek has a solid lead here, with its dominance of the cellular modem/device market, but expect to see solutions pairing MediaTek, Quantenna and Broadcom modems with other router SoCs, whereas Qualcomm will probably provide fully integrated solutions out of the box. We even expect to see Western Digital add more high-end devices to its current mobile line-up.

Another paradigm change is the use of the OpenWrt OS, as more manufacturers embrace its standards and compatibility. Also expect to see greater emphasis on security, in parallel with hardware features designed to support "smart router" functions (i.e., remote administration via smartphone or Web apps, which at the moment, is a feature solely up to individual router manufacturers to implement).

On the negative side, the 2.4 and 5GHz bands used for Wi-Fi are becoming more crowded. Interference from multiple devices on these bands, especially in public and enterprise Wi-Fi spaces, means interference and higher error rates, all of which serve to slow down individual connections regardless of actual hardware capabilities. As this is somewhat of a physics-imposed limitation, expect to see active workarounds that include the shunting of smart-home-appliance and IoT connections to other transmission bands.

The 802.11ah extension to the 802.11 Wi-Fi standard allows the use of sub-1GHz bands for Wi-Fi communications, and will be up for approval in its entirety in March 2016. A sub-component to the “ah” extension, the “HaLow” standard operates on the 900MHz band and was recently approved by the Wi-Fi Alliance. It allows for low-power and high-obstacle-penetration operations.

The next iteration of conventional 2.4GHz and 5GHz Wi-Fi is expected to be the 802.11ax standard, still in early stages of development, but which promises 10 Gb/s speeds. Finally, expect further work on all the other iterations of 802.11 standards that utilize bands other than the 2.4GHz and 5GHz, specifically 802.11af that uses the unused TV bands (UHF/VHF white-space spectrum) between 54 and 790MHz.

Follow Gene Fabron @fabrongene. Follow us on Facebook, Google+, RSS, Twitter and YouTube

MORE: Wireless Routers 101
MORE: All Networking ContentMORE: Networking in the Forums

Gene Fabron is a Contributing Writer for Tom's Hardware. Follow her on Twitter. Follow us on Facebook, Google+, RSS, Twitter and YouTube.

TP-Link marked its entry into the Smart Home market with the SR20 Smart Home Router. Potential buyers looking to set up a Smart Home network might want to consider what TP-Link has to offer. The SR20 is a fully-featured AC1900 router that acts as a central hub for connected devices throughout the home network.

On the router side, the TP-Link SR20 is equipped with a 1GHz dual-core processor and delivers wireless speeds of up to 1300Mbps on 5GHz and 600Mbps on 2.4GHz. Beamforming technology allows the SR20 to provide stable coverage over longer distances. Further connectivity is provided through four Gigabit Ethernet ports. Storage is handled by a built-in SD card slot, and file sharing is handled via one USB 3.0 and one USB 2.0 port.

TP-Link's SR20 fulfills its place as a Smart Home Router by acting as the focal point for connected IoT (Internet of Things) sensors. In addition to controlling sensors through the SR20's built-in touchscreen, the sensors may be controlled through a connected smartphone or tablet using the Kasa app. An example I saw during TP-Link's demo at CES were LED lamps changing color in response to a water leak. Another scenario users may configure is having their thermostat raise the temperature at a certain time so that the user returns to a nice, toasty home.

The SR20 achieves all of this via compatibility with ZigBee and Z-Wave. Since such a large ecosystem has spawned from ZigBee and Z-Wave, the SR20's functionality as a Smart Home Router is widespread.

Smart Home technology is an evolving market, and it's great to see a unified product that can govern all aspects of a Smart Home, but TP-Link isn't the first company to do this. Given time and more innovation from Smart Home developers, I expect other router vendors to provide their own solutions as well.

Alexander Quejado is an Associate Contributing Writer for Tom’s Hardware and Tom’s IT Pro. Follow Alexander Quejado on Twitter. Follow us on Facebook, Google+, RSS, Twitter and YouTube.

Securifi, a router vendor known for its successful Kickstarter campaign, has announced the availability of the Almond 3, an update to the previously released Almond 2015. The newly released Almond 3, like its Almond + and Almond 2015 counterparts, acts as both a router and a Smart Home hub.

The Securifi Almond 3 is an excellent router on its own; it is an AC1200 (867 Mbps on 5 GHz and 300 Mbps on 2.4 GHz), offering a range of up to 5000 square feet. The Almond 3 is equipped with two 2x2 antennas for coverage and one USB 2.0 port for file sharing or USB dongles for Z-Wave/Bluetooth configurations. Also, this wouldn’t be an Almond product if it didn’t have a touch screen. Securifi also retains its distinction of being the first vendor to put a touch screen on a router by slapping on a 2.8 inch touch display.

The Almond 3 is a more than just a router. It is a centralized hub capable of managing a wide variety of devices thanks to its built-in Zigbee radio and support for Z-Wave and Bluetooth. Since the ZigBee and Z-Wave ecosystems are incredibly community driven, the Almond 3’s functionality is nearly limitless; a list of compatible sensors is available on Securifi’s website. Aside from the touch screen, the Almond 3 may configure sensor rules with a smartphone or tablet using the Almond mobile app.

Here a couple of examples of how a unified Smart Home hub can be configured. Securifi, at the very least, recommends a home security system, so the Almond 3 features a programmable siren. Further steps to expand your home security system might be installing ZigBee sensors that detect whether doors or windows are open, or turning on a camera when someone approaches home entrances. Since the Almond 3 is a router, users may even configure their security system to shut off as soon as the router detects the user’s. This can be done without even taking one’s phone out of their pocket, of course. Beyond security, someone may manage thermostats, smoke alarms, lights, etc. from the router.

The Securifi is currently accepting preorders for the Almond 3, which will retail for $119.

Alexander Quejado is an Associate Contributing Writer for Tom’s Hardware and Tom’s IT Pro. Follow Alexander Quejado on Twitter. Follow us on Facebook, Google+, RSS, Twitter and YouTube.

With consumer-level virtual reality just around the corner, we're seeing more and more companies pop up in this space with innovative ideas. Several of these companies are trying to make it as easy as possible to share content in the virtual world, but they all have a different vision about how to do that.

Vicarious is one of those companies. I recently had a chance to speak with J.M. Yujuico (pronounced You-wee-ko) and Alexander Chu, the co-founders of Vicarious, to learn more about the company and their vision for the platform that their company is building.

Vicarious is a startup based in San Francisco. The company is building a platform that will let creators publish any content they desire and have it available to consume in virtual reality. The founders believe that much of that content will actually be 2D images and video, and they are creating tools that will make it easy for people to share that content in creative ways on VR devices.

The Vicarious platform will consist of three components: the Vicarious Toolset, Vicarious Viewer and the Companion App.

The Vicarious Toolset will enable all sorts of creative ways to display content. Creators will be able to edit and customize these experiences from within the VR application. The platform is being designed to support many different types of content, including 2D, 180- and 360-degree video and photos, 3D objects and animations, audio, and even text. The plan is to make the platform completely source agnostic.

Yujuico and Chu believe that the biggest hurdle to creating VR content will be access to recording equipment in the early days of consumer virtual reality. Their hope is that providing the ability to share any content in a VR environment will create a wealth of content available for consumption in VR.

To view the content created for Vicarious, there will be a Vicarious Viewer application. This app will be available for GearVR and Google Cardboard, but Cardboard users will be limited to viewing only, while GearVR users will be offered more function such as customizing your own showcase for your content. The company plans to support PC-based headsets such as the Oculus Rift and HTC Vive in the future, but currently, development is focused on creating the platform and rolling out. Further support will come later.

Vicarious is also developing a Companion App. This application will run on your smartphone, not in VR, and let you edit your content page. You can add video and images from the app to your VR experience when you're out and about without a headset.

The Vicarious platform is still in its infancy. The company has not yet announced a beta release, but it expects to launch the product in the first quarter of 2016. I was told there will be some announcements about Vicarious before the end of the year.

Follow Kevin Carbotte @pumcypuhoy. Follow us on Facebook, Google+, RSS, Twitter and YouTube.

The Thread Group, which develops the Thread mesh networking protocol for connected homes, launched its certification, and more than 30 products have already been submitted for testing. The Thread specification version 1.0 was released earlier this summer.

“Having already made the Thread 1.0 technical specification available to members under Thread's RAND-RF (royalty-free) licensing terms, today we are taking another major step in bringing Thread certified products to market, and ensuring that they are tested for quality, security and interoperability," said Chris Boross, president, Thread Group. “This is great news for our 220 member companies, who are anxious to start using Thread in their products."

The group was started by Yale Security, Silicon Labs, Samsung Electronics, Nest Labs, Freescale Semiconductor, Big Ass Fans and ARM in July 2014. Since then, Qualcomm has become one of the major sponsors as well, and there are already tens of other contributing companies such as Intel, Imagination, Microsoft, LG, and many more affiliate companies, too.

The Thread protocol is based on another low-power wireless standard called '6LoWPAN' (IPv6 over Low-power Wireless Personal Area Networks) and will work on top of 802.15.4 hardware, currently already used by some Nest and Zigbee products.

The main advantages of Thread over other protocols is that the mesh network it can create between multiple smart devices at home is highly resilient, as there's no single point of failure, it's secure by default, and it's also IP-based.

Because it's IP-based, any of the up to 250 devices in the network can also connect directly to the Internet. Although, with all the warnings about poor IoT security, it's probably not the best idea to connect most smart devices to the Internet, if it can be helped. The main idea behind Thread remains that it's a mesh networking protocol, which can connect hundreds of devices to each other, locally.

Now that the full specification was released and the Thread Group is starting to certify devices for the protocol, we should start seeing Thread-enabled devices in the market in a matter of months. Certified software stacks from ARM, Freescale and Silicon Labs should already be available this month.

Thread's board of directors was also recently joined by Osram, a global company that sells lighting products and is presumably also interested in making Thread-enabled smart lightbulbs in the future.

"Being a part of the Thread Group will significantly strengthen our effort to drive innovative and sustainable lighting solutions that can enhance any environment," said Timon Rupp, CEO, OSRAM Lightify. "It provides an unprecedented opportunity for OSRAM to collaborate with other visionaries and be on the front lines of driving smart solutions beyond energy savings," said Geert van der Meer, senior vice president, OSRAM Digital Systems.

The Thread products that will be certified will be tested for proper device behavior, networking functionality, security and operation in the Thread network. They may also come with a "Built on Thread" or "Thread Certified Component" logo as a guarantee to consumers that the Thread functionality is working as it should for the home automation devices they buy.

The Thread Group also wants to see what kind of innovative ideas for the connected homes startups can come up with, and for that purpose it has created the Innovation Enabler Program. Every quarter, the group will announce some winners who will receive up to 18 months of Thread membership. For Q3, the winners are Centero, an IoT technology and integration services company, and iSocket, a developer of smart plugs for the connected home.

Other companies, who want to participate in the program, can visit the Innovation Program site for more information.

______________________________________________________________________

Lucian Armasu joined Tom’s Hardware in early 2014. He writes news stories on mobile, chipsets, security, privacy, and anything else that might be of interest to him from the technology world. Outside of Tom’s Hardware, he dreams of becoming an entrepreneur.

You can follow him at @lucian_armasu. Follow us on Facebook, Google+, RSS, Twitter and YouTube.

About a year ago (July last year), Nest, along with several other companies, joined together to form the Thread Group consortium, in an effort to bring a more secure low-power networking protocol for IoT devices.

The founding members included Yale Security, Silicon Labs, Samsung Electronics, Nest Labs, Freescale Semiconductor, Big Ass Fans and ARM, but now the group contains over 120 new members.

Thread mesh networks could be used in a smart connected home for appliances, access control, climate control, energy management, lighting, safety or security. The protocol has many advantages including being low-power as well as it's an open protocol, resilient, IPv6-based and secure by default. It's also designed to provide a fast time to market for interested IoT device makers.

The Thread mesh networks will be built on top of 802.15.4 hardware, which is already on the market in some Nest devices and is what Zigbee, a previously competing wireless protocol for home automation, used as well. However, Zigbee will also be built on top of Thread at the application layer, so in the future it should become more of a complementary product than a competitor.

Thread uses the 6LoWPAN protocol, on top of which it builds the mesh network. The 6LoWPAN acronym stands for "IPv6 over Low-power Wireless Personal Area Networks." That means the Thread networks are IP-based, and the devices can also connect directly to the Internet, not just to each other. This is an advantage the protocol has over other wireless protocols such as Bluetooth and Zigbee. It's also likely the main reason why Zigbee will be built on top of Thread in the future.

Compared to competitors, a Thread mesh network is also more resilient and can extend the range of a home network of smart devices. If one node fails to connect to the network, other devices in the network will still be able to connect to each other. This is the advantage of a mesh network over more centralized approaches.

Bluetooth, for instance, has an effective range of 50 meters, and the devices can only connect in pairs of two. This limits the usefulness of Bluetooth in smart homes because it makes managing multiple smart devices much more difficult. On the other hand, a Thread mesh network requires no maintenance after the initial setup for the new devices that join it.

Over 250 devices can be connected to each other in a Thread mesh network as long as any two of those devices have a reasonable range between them. Therefore, a Thread network could easily cover a large house or property without any signal loss. Where two Thread nodes are too far away from each other, the range can be extended with "Thread routers."

The Thread networks will have a bandwidth of about 250 Kbps, which isn't enough to transfer large files between the devices, but it can still enable the type of communications sensors can have with each other. The low bandwidth is a compromise that had to be made to keep the Thread-enabled devices low-power and last not days, but years, on small batteries. The latency is less than 100ms for typical interactions.

Many security experts worry that IoT (nicknamed the "Internet of Threats" by Kaspersky Lab's founder) is going to make hacking exponentially more common once many people begin buying insecure smart devices for their homes. That's why Thread comes with built-in security that's enabled by default and mandatory for all devices.

Users will have to authorize any Thread-enabled devices before they are allowed on their home networks. To communicate with each other, the devices will also have to recognize each others' MAC addresses, which should make it harder for other unauthorized devices to access the network. The communications between authorized devices will be encrypted with DTLS (Datagram Transport Layer Security), an encryption protocol designed to prevent tampering and message forgery.

The Thread Group has recently finished the Thread protocol specification, and it will begin certifying products this fall. The first Thread-enabled devices for connected homes are expected to come out next year. As original founders, Google's Nest and Samsung, helped by its recent SmartThings acquisition, could be among the first companies to ship Thread-enabled products for smart homes.

Follow us @tomshardware, on Facebook and on Google+.

Remember the Skyrim mod created by Alexander J. Velicky that landed news coverage in July? Back then this monster of a project, called Falskaar, just launched for free to Skyrim players, and promised more than 25 hours of new content. That included a new land mass a third the size of the original game, new characters, new voices, an original soundtrack and dozens of quests. Over 100 people contributed to this mod in some form which Velicky took and plugged into Falskaar himself. However, he produced almost all of the content on his own.

Crazy, we know, but the mod was more than content generated by a faithful fan; it was basically a demo to show his modding and management skills so that perhaps Bethesda would take notice and give him a job. Bethesda took notice alright, but it wasn't Bethesda that finally caved in to his talents, but none other than Bungie, the studio behind the Halo series and the upcoming Destiny game.

In an open letter posted on Bethesda's forums, Velicky reports that he has accepted an Associate Designer position at Bungie. He then thanks the community for supporting the development of his massive mod, and even thanks those who did the "simple" things such as liking a video, endorsing a mod, providing feedback and so on. He also provides some advice for budding developers.

"Never be afraid to try," he writes. "I set my sights on a professional design job pretty early, I lowered my head, charged forward, and rarely looked back. Of course, I ensured what I was doing had a reasonable chance for success from time to time. But the most surprising of all, is who I've ended up with."

"I applied to many companies, and Bungie was in my, 'Huge company that will completely ignore me' category," he continues. "Well, they didn't and look what it got me. Bungie is an awesome company with an amazing team, and I'm very lucky that they've decided to give me a chance! Never be afraid to try. I spent the time it took to apply and the rewards are proving to be greater than I could have possibly imagined."

The 19-year-old developer previously said he graduated from high school over a year ago. Instead of going to design school, he made Bethesda's Creation Kit his own classroom. He then spent 2,000 hours over the last year building and populating Falskaar. That's a lot of dedicated time, especially when he's done with high school and still living at home.

"[My dad] was incredibly supportive and allowed me to live here, paying for living expenses and charging no rent," Velicky said. "I was able to not go to school and not have a day job. Meaning, more or less, that Falskaar was my day job."

Now he works at Bungie. Dad must be really proud!

Phoronix reports that an Android version of the open-source Wine software – which is used to run Windows-based apps on POSIX-compliant operating systems like Linux, Mac OSX and BSD – is currently in development.

The Android version was briefly demonstrated by Wine's original developer Alexandre Julliard after the FOSDEM 2013 talk in Brussels on Monday. Phoronix reports that CodeWeavers still employs Julliard to work on the software due to their Wine-based CrossOver commercial software for Linux and Mac OS X.

Phoronix said the performance of Wine for Android during the demo was "horrendously slow" because Julliard didn't use an actual Android device, but rather an emulated Android environment running on an Apple MacBook. It's also an active work-in-progress, so performance is expected to be somewhat slow for now.

"While Wine is coming to ARM and there's quite a lot of interest there, CodeWeavers is quite interested and hopeful for the success of Intel x86 Atom CPUs for tablets," Phoronix said. "If Android gains traction on x86-based tablets and other mobile devices, CodeWeavers has a lot of commercial opportunities for pushing the running of Windows software on Android."

According to the Wine website, the software doesn't simulate internal Windows logic like a virtual machine or emulator. Instead, Wine translates Windows API calls into POSIX calls on-the-fly. This supposedly eliminates the performance and memory penalties associated with VMs and emulators, allowing the user to "cleanly" integrate Windows applications into a non-Windows desktop.

Wine originally began in 1993 as a way to run Windows 3.1 programs on Linux. It took 15 years before Wine finally reached v1.0, the first stable release, in 2008.

_{Contact Us for News Tips, Corrections and Feedback}

EVGA Overclocks The GeForce GTX 460

When the Radeon HD 6800-series cards were launched in October, Nvidia representatives wanted us to include EVGA's GTX460 FTW in the benchmarks to represent the Radeon HD 6870’s competitive landscape. We didn’t go that route because the idea of benchmarking a GeForce GTX 460 with the highest available factory overclock against AMD's standard Radeon HD 6870 didn’t sit very well with us. Although EVGA’s highly overclocked GeForce GTX 460 is a perfectly valid product, it’s not a reference design, and is only offered by one company, while many manufacturers produce cards based on the Radeon HD 6870 reference standard.

With the Radeon HD 6800-series launch behind us, though, we’re certainly interested in seeing what EVGA’s top-of-the-line GeForce GTX 460 can do, and how it compares to its competition, Radeon- and GeForce-based cards alike.

From what we’ve seen, EVGA’s card boasts the highest factory overclock of any GeForce GTX 460 1 GB available to date. With 850/1700 MHz core/shader speeds and a 1000 MHz GDDR5 memory clock, this card has a significant 175/350 MHz core/shader and 100 MHz memory advantage over Nvidia's original design.

This overclock looks especially impressive when you compare stats against the reference GeForce GTX 470. EVGA’s GTX460 FTW has the potential to perform about 5% more shader operations per second than a reference GeForce GTX 470, and practical memory bandwidth is brought within 5% of the GF100-equipped card. The overclocked core and shader speeds push raster performance into an even higher realm. This suggests that EVGA’s factory overclocked GeForce GTX 460 should perform in the neighborhood of the GeForce GTX 470, and maybe even beat it in some situations. It’s no wonder Nvidia partially crippled the GF104 GPU and kept its reference clocks lower than they probably could have been—the company didn’t want its mid-range offering showing up in high-end products.

If you were expecting a magnum-force custom-cooling solution, prepare to be surprised. EVGA’s wunderkard humbly makes use of the reference cooler, tastefully arrayed in EVGA colors.

As far as we can tell from the back of the board, the PCB is reference fare, too. The only hardware upgrade that EVGA mentions in its marketing materials is an upgraded MOSFET heat sink.

This card requires two PCIe power connectors, just like every other GeForce GTX 460 we’ve seen.

The card’s outputs are also standard-issue GeForce GTX 460 equipment: two dual-link DVI outputs, complemented by a single mini-HDMI output.

The bundle includes a DVI-to-VGA adapter, two Molex-to-PCIe power-connector adapters, a user guide, a driver CD, and what appears to be a bumper sticker.

Test System And Benchmarks

We’re including a number of cards in our benchmarks. From AMD, we have the Radeon HD 6850 and 6870. From Nvidia, the GeForce GTX 460 and 470, in addition to EVGA’s factory overclocked GTX460 FTW, of course.

As described in our Radeon HD 6800 launch article, the majority of GeForce GTX 460 cards on the market are at least slightly overclocked beyond the reference specification. To make sure the benchmarks reflect what's available for purchase, we averaged out the clocks for the seven cheapest models we could find for a fair representation of the GeForce GTX 460 playing field. As a result, we are benchmarking the GeForce GTX 460 1 GB at 708/1416 MHz core/shader speeds for our tests, a 33 MHz increase over the original reference core clock. This increase is so low that, frankly, it will probably have no measurable impact on our results. It does seem like the fair thing to do, though. The memory runs at the reference 900 MHz GDDR5 clock, as even overclocked models tend to leave memory at reference speeds.

It's also worth noting that most Radeon HD 6870s currently available ship at reference clock rates. There are a handful of 13 and 20 MHz speed bumps, and a couple of 40 MHz increases that command higher prices. However, for the folks ready to hit the comments section with, "Sure, compare overclocked Nvidia hardware to stock AMD cards," bear in mind that this is representative of today's retail landscape.

Benchmark Results: 3DMark Vantage And Unigine Heaven

With the performance preset at 1280x1024, the EVGA GTX460 FTW’s high overclock kicks into gear and delivers an impressive result, beating out even the Radeon HD 6870 and GeForce GTX 470. This is somewhat optimistic, since our real-world tests happen at higher resolutions, but it does show us the potential of EVGA’s card.

Without tessellation enabled, the EVGA GTX460 FTW leads the pack, but once the feature is turned on, the card performs on par with the GeForce GTX 470.

Benchmark Results: Metro 2033 (DX11)

EVGA’s overclocked GeForce GTX 460 bests the Radeon HD 6870 here, and almost catches up with the GeForce GTX 470.

Benchmark Results: Lost Planet 2 (DX11)

In Lost Planet 2, the EVGA card keeps up with the GeForce GTX 470 at 1680x1050, but slows down at 1920x1080. At 2560x1600, all three of the powerful contenders are stuck in a dead heat.

Benchmark Results: Aliens Vs. Predator (DX11)

In this title, the Radeon HD 6870 has a slight advantage in average frame rates, but a notable minimum frame rate lead. The EVGA GTX460 FTW performs closely to the GeForce GTX 470.

Benchmark Results: Battlefield Bad Company 2 (DX11)

Once again, the overclocked EVGA card performs closely to cards in the $250 range. Though the EVGA board itself costs $240, it drops to $220 with rebate.

Game Benchmarks: StarCraft 2 (DX9)

This CPU-limited game shows little difference between graphics hardware.

Benchmark Results: Civilization 5 (DX11)

Civilization V has two distinct areas where it can be tested: diplomacy and the map. The diplomatic view involves beautifully rendered versions of famous leaders like Alexander the Great and the Pharaoh Ramesses, while the map view involves representations of military forces on well-rendered terrain, using DirectX 11's tessellation features to help improve performance, rather than improving realism. We start with the leader benchmark, a test that does give us some comparative information about the products we're testing.

The results are so astronomical that any advantage here is entirely theoretical.

The GeForce cards perform well versus their Radeon counterparts, but all of the frame rates are quite high here.

Benchmark Results: DiRT 2 (DX11)

The overclocked GeForce GTX 460 does very well here, beating the GeForce GTX 470 across the benchmarks. The playing field is tight, though.

Benchmark Results: Just Cause 2 (DX11)

Once again, the overclock helps the GeForce GTX 460 perform closely to the GeForce GTX 470.

Benchmark Results: DiRT 2 With Anti-Aliasing

In DiRT 2, the EVGA GTX460 FTW performs excellently whether anti-aliasing (AA) is applied or not. When AA is disabled, the card easily surpasses the other options.

Benchmark Results: Just Cause 2 With Anti-Aliasing

In this game, the EVGA card also performs excellently when AA is enabled. The performance penalty it suffers is relatively small.

Power Usage And Temperature Benchmarks

Despite its GeForce GTX 470-class performance, the EVGA GTX460 FTW is using a lot less power under load, although the Radeon HD 6870 manages to achieve excellent power efficiency in comparison to both.

The cooler might not be fancy, but it does a great job. It’s interesting to note that, at idle, we saw the EVGA clocks go as low as 50 MHz core and 67 MHz memory.

The EVGA card even manages to perform relatively quietly compared to the higher-end GF100-driven GTX 470.

Conclusion: EVGA GTX460 For The Win!

If nothing else, EVGA’s GTX460 FTW shows us what the GF104 GPU is capable of. It’s crystal clear why the GeForce GTX 460 was given a relatively conservative reference core and shader clock relative to its limits, along with a single disabled Shader Multiprocessor. After all, it can go toe-to-toe with the GeForce GTX 470 when everything is dialed up. It'd be nice if Nvidia released a card sporting GF104 GPU and all eight streaming multiprocessors enabled, just to see what it can do.

Despite the excellent factory overclock, value always comes down to a balance between price and performance. It’s been a couple of months since the Radeon HD 6800 launch, so let’s have a look at the current market, as represented on Newegg. At the time of writing, most of the Radeon HD 6850s hover around the $190 range. The GeForce GTX 460 1 GB can be found for $180-250. EVGA’s overclocked GTX460 FTW sells for $230, a $50 premium over the cheapest GeForce GTX 460 1 GB available. The Radeon HD 6870 is readily available around $240. Finally, the GeForce GTX 470 sells for around $250.

Here’s where the decision gets dicey: do you want an EVGA GTX460 FTW, a Radeon HD 6870, or a GeForce GTX 470? With a $10 spread between these options, it’s very hard to call a clear winner here. Frankly, any one of these cards is going to provide excellent service. A typical user is never going to regret having chosen one of these options over the other, as performance is practically identical and they all have their own strengths and weaknesses once you move past the average frame rates. The GeForce cards have CUDA and 3D Vision, while the Radeon boards have Eyefinity. Pick the value-adders that're important to you.

The bottom line is that EVGA’s GTX460 FTW offers something that I’ve never before seen in a card with an extreme factory overclock: excellent performance relative to its price. With GeForce GTX 470-class gaming ability combined with lower power usage, noise, and heat, EVGA’s option truly lives up to it's name and surpasses other GeForce GTX 460 cards for the win.