AI Intel

Claude Code v2.1.287 introduces "Claude Mods" allowing deeper plugin behavior modification and a "You should know" feature where a side agent flags p…

llama.cpp version b11414 resolves a critical Vulkan bug concerning stale prealloc_y reuse across flash attention and softmax operations.

Spira Maxima is a new video model that transforms scripts into engaging, viral social media videos.

Ollama v0.40.0 automatically runs models on MLX by default on Apple Silicon devices, significantly enhancing performance.

vLLM v0.31.0 features 717 commits and significantly improves DeepSeek-V4.1-Flash performance using FlashMLA mega attention with NVFP4 compressed KV c…

A Qwen3.8-Flash-Next model hallucinated a signed Alibaba Cloud URL during product research, prompting questions about its unusual behavior.

devpit, a new tool on Product Hunt, offers a native control room experience for managing Claude Code agents.

Reflection AI is poised to release a new "strong" US open-weight model, aiming to compete with leading models like DeepSeek and Qwen.

Clef Flash demonstrates the ability to play the game Snake in real-time on an RTX 5080 using simple instructions, requiring no prior training or game…

llama.cpp now supports decision models via its new `/v1/systemone` endpoint, enabling local, efficient, and private Jev-style inference.