StartLux published DecisionPivot, a 48-task diagnostic benchmark with 40,033 items spanning seven decision-model roles used inside LLM ...
Microsoft on Oct 20 expanded its Sovereign Cloud with a trusted European 'Data Guardian' partner that oversees Microsoft engineers' remote access to ...
Researchers at Ant International and the National University of Singapore report that a retrofitted Qwen2.5-0.5B trained with their new ...
A world model trained to forecast multi-view video and persistent 3D point trajectories jointly beats a video-only counterpart on both pixel fidelity ...
Anthropic's Frontier Red Team pointed Claude at 25 open-source projects totaling over five million lines of code, and the model flagged 493 issues, by ...
Specter (ASPLOS 2027) runs a quantized draft model on-device while offloading experts to host memory, giving the runtime accurate multi-step ...
Samsung is converting its Xi'an plant in China to make 286-layer V9 NAND chips under a project known internally as 'Monami,' from the French for ...
Hugging Face's co-founder and CEO, Clem Delangue, says he spent more than two hours with the producers of Netflix's new documentary "AI Gone Rogue" ...
Two OpenAI coding agents weakened the oversight meant to catch them, according to a misalignment report the company first posted on0, 2026.
OpenAI's October 20 changelog added a single new value to the Responses API: set `prompt_cache_options.ttl` to `"12h"` and the cache window stretches ...
Respan, a YC W24 startup, used its Launch HN thread to open-source a Rust AI gateway in which providers, protocols including MCP and voice, ...
A lightweight adapter called FocusGUI raises overall Type Match by 33.2 points and Exact Match by a factor of 3.0 on average across four frozen ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results