The world of artificial intelligence (AI) is growing at an exponential rate, with a broad range of topics spanning from the latest CPUs, GPUs, ASICs and FPGAs that run modern AI workloads, along with the different types of AI usages, such as the different types of large language models (LLM) and how they are trained and then used for inference workloads. Here you'll find Tom's Hardware's leading coverage of all things AI.
Latest about Artificial Intelligence
Anthropic's latest AI model identifies 'thousands of zero-day vulnerabilities' in 'every major operating system and every major web browser'
Anthropic holds back its most advanced model yet to allow companies and institutions to prepare.
Half of planned US data center builds have been delayed or canceled, growth limited by shortages of power infrastructure and parts from China
As cloud giants plan to spend $650 billion on AI infrastructure this year, the availability of power infrastructure components has become a significant obstacle to deploying AI data centers.
Microsoft says Copilot is for entertainment purposes only, not serious use
Microsoft Copilot's terms of service explicitly say that it's "for entertainment purposes only," but its marketing says otherwise.
Iran claims it has hit Oracle data center in Dubai, Amazon data center in Bahrain
Dubai denies, Bahrain confirms hit on Batelco, an AWS partner.
The largest programming community on Reddit has just banned all content related to LLMs
That's why AI as a whole isn't banned.
Meta to fund seven new natural gas power plants to fuel AI data centers — Entergy partnership to deliver 7 gigawatts of power for Louisiana AI facility
Meta is paying for the construction of seven new natural gas plants to supply its largest data center, massively expanding the fossil fuel infrastructure underpinning its AI buildout.
US judge sides with Anthropic, says company supply chain risk branding over Pentagon disagreement 'Orwellian'
A U.S. federal judge ruled that the Pentagon cannot brand Anthropic as a "supply chain risk" simply because it refused to give in to its demands.
Google's TurboQuant reduces AI LLM cache memory capacity requirements by at least six times
In benchmarks on Nvidia H100 GPUs, 4-bit TurboQuant delivered up to an eight-times performance increase in computing attention logits compared to unquantized 32-bit keys.
