LLMs

Satya Nadella revealed that AI writes as much as 20% to 30% of the code in Microsoft's repositories and projects.

Nvidia has developed a new plug-in builder for G-Assist, enabling it to integrate with various large language models (LLMs) and software to enhance functionality.

OpenAI seems to have inadvertently tuned ChatGPT into a potent geo-guesser.

AMD's RDNA 4 workstation GPUs are rumored for announcement at Computex next month.

FramePack allows you to generate a one-minute clip at 30 FPS using a 13-billion parameter model on a 6GB graphics card.

Microsoft researchers developed a 1-bit AI model that's efficient enough to run on traditional CPUs without needing specialized chips like NPUs or GPUs.

An in-depth review compares both 8GB and 16GB VRAM models of the RTX 5060 Ti. For many games, the 8GB is just as fast, but there are also plenty of instances where it can run out of VRAM.

Amuse 3.0 is finally here, offering Stable Diffusion models that are fine-tuned for AMD hardware.