On MBP M3 Max benchmarks, the gap between three Qwen3.6 variants and gpt-oss:120b was revealed.
Comparing 10 image models on an M1 Max 64GB revealed differences in non-English cultural accuracy.
The Plano 0.4.22 update adds a TUI feature for monitoring Claude Code's costs and adaptive routing.
Hiring bias was confirmed in which LLMs favor resumes they themselves generated.
A regression has surfaced where VS Code adds `Co-authored-by: Copilot` regardless of whether Copilot was actually used.
Context Mode, built for AI coding agents, has been introduced as reducing context usage by 98%.
In a Harvard study, OpenAI o1 achieved 67% accuracy in initial emergency room diagnosis.
Penn improved memory, training time, and stability for inverse PDEs using Mollifier Layers.
The Blender MCP server executes LLM-generated code without safeguards, raising the risk of data leaks.
The latest llama.cpp build significantly improved the RTX6k Qwen3.5 benchmark.
An `.xcstrings` localizer plugin for Claude Code has been released.
Heard has been released as a voice guidance tool for Claude Code and Codex.
By adding forward-derived fast memory on top of a frozen Pythia-70M, one-shot symbolic recall was attempted.
Vex, an open-source tool that supports data migration between vector databases using different standards, has been released.
The OpenAI privacy-filter has been retrained on Nemotron PII data to support 55 fine-grained PII labels.
Elon Musk has launched a full-fledged legal battle, claiming OpenAI abandoned its nonprofit mission.
Qwen has released a Top-K SAE interpretability module and checkpoints for 64 layers for Qwen3.5-27B.
The CppStudio framework has been released to help AI coding agents perform specialized GPU programming.
Proposes an architecture that separates the execution loop from the sandbox to improve the security and scalability of AI agents.
Unitree's G1 humanoid robot successfully demonstrated ice skating and roller blading through sophisticated balance control.
The VS Code Git extension has changed so that it attaches `Co-authored-by` to commits regardless of whether Copilot was used.
OpenAI has enabled marketing cookies by default for free ChatGPT users in the US.
It explains that schema validation raised the CoT compliance rate from 9.91% to 100%.
An 826,985-parameter Transformer built in C++17 achieved a validation loss of 1.6371 after training for 76.2 minutes on a single CPU core.