Daily AI news, sourcedUpdates / All
Updates Official · 987 items · newest issue first · page 26 of 50
AreaAllModelsAgents & codingApps & productsChips & infraResearchPricing & limitsBusiness & fundingPolicy & safetyOther
- Z.ai unveils GLM-Legal, a GLM-5.3-Flash-based legal agent model scoring 19.2% all-pass on Harvey LABOfficialZ.aiOct 13Models
- Researchers release U-Space, a label-free way to read LLM uncertainty from hidden states in a single pass
- Google DeepMind's AlphaProof Nexus paper, which solved 9 open Erdos problems in Lean, is published in Science
- Microsoft will scale Azure Maia to multi-gigawatt deployments from 2027, making it the primary inference system for its products
- NVIDIA contributes Vera Rubin rack and compute tray designs to OCP and open-sources new DSX AI factory software
- Microsoft researchers' Fork-and-Flush clears agent chat history but keeps the workspace, lifting long-run resultsOfficialarXiv (Microsoft Research)Oct 13Research
- shad0w learns a local LLM's classification answers and serves the repeats in microseconds
- LSearch is an open-source Spotlight-style semantic search launcher for Windows with a local fixer model
- LocalLightChat, a lightweight Open WebUI alternative, adds agents, MCP, RAG and ComfyUI image editing
- NuMind open-sources a vLLM extension that returns a served LLM's last hidden state as an embedding
- Gewell 0.2 inference engine adds Gemma 4 26B A4B and multimodal EmbeddingGemma 2 servingOfficialGitHub leDissolution/gewellOct 12Models
- Qwen introduces QwenGyre, an elastic RL framework for training agents on extremely long tasks
- BrokenML benchmark tests how much cluster access AI agents need to fix broken ML jobs
- Graphsignal adds graphsignal-top, a live top-style terminal view of inference GPUs and kernels
- Atticus is an open-source hold-to-talk voice assistant for macOS that runs on Claude Code
- TinyDetect runs AI text and image detectors fully in the browser
- CL4SM opens a call for contributions on constrained learning for scientific modelling ahead of NeurIPS Paris
- Matthew Schwartz's BootLoops toolkit lets LLMs carry exact methods into 22 scientific fields
- whatmodel.sh ranks the best AI models your own machine can run
- strix-llama rdna35-perf branch speeds Strix Halo prompt processing vs Vulkan and stock ROCm