inferenceoptimizationopen-source
ik_llama.cpp Delivers 26x Faster Prompt Processing for Qwen 3.5
A new optimized C++ inference engine achieves 26x speedup on Qwen 3.5 prompt processing, a major win for local AI deployment.
2 min
What's New in AI
Releases, research, trends — everything important in plain language
Page 3 of 3
moonshotcursorlicensing
Moonshot Confirms Cursor Was Authorized to Use Its Model
Moonshot AI clarifies the Cursor Composer controversy, confirming the partnership was legitimate. Here's what it means for AI model licensing.
2 min
nvidianemotronopen-source
Why NVIDIA's Nemotron Cascade Is the Most Efficient Reasoning Model Yet
Nemotron-Cascade 2 achieves Gold Medal math performance with just 3B active parameters. NVIDIA's open model is redefining what small models can do.
7 min
openairesearchsafety
What Happens When AI Gets Infinite Compute? OpenAI's Disturbing Discovery
OpenAI's research team reveals that models go 'insane' when given unlimited reasoning tokens — raising new questions about AI scaling and safety.
7 min