Technology & AI
ProjAgent hits 41% on REPOCOD with procedural code retrieval
ProjAgent adds a new retrieval signal that finds code doing the same job under different names, hitting 41.14% Pass@1 on the REPOCOD benchmark for repository-le
Technology & AI
Daily AI and technology news decoded in plain English — models, chips, agents and research, and what each development actually means for you and your business.
418 stories
Technology & AI
Quantized LLMs behave differently despite matching accuracy
A July 2026 preprint finds models compressed to lower precision diverge in which questions they get right, even when overall accuracy holds steady — exposing a
Technology & AI
SLORR cuts LLM training overhead below 1%
A new arXiv preprint called SLORR adds under 1% training overhead in LLM pretraining while making models more compressible, without SVDs or architecture changes
Technology & AI
OpenAI's ChatGPT Work agent runs across your apps for hours
OpenAI's new ChatGPT Work agent promises to turn goals into finished work by acting across your apps and files for hours — but key details remain undisclosed.
Technology & AI
GPT-5.6 becomes preferred model in Microsoft 365 Copilot
OpenAI's GPT-5.6 is now the preferred model powering Microsoft 365 Copilot across Word, Excel and PowerPoint, affecting millions of daily office users worldwide
Technology & AI
OpenAI GPT-5.5 Bio Bug Bounty targets biological jailbreaks
OpenAI's GPT-5.5 Bio Bug Bounty invites researchers to test universal jailbreaks for biological risks in a model built for autonomous, multi-tool work.
Technology & AI
OpenAI GPT-5.6 arrives with Sol, Terra and Luna trio
OpenAI's GPT-5.6 launches a three-model lineup — Sol, Terra and Luna — promising token efficiency and frontend gains, but with no independent benchmarks yet.
Technology & AI
Strong privacy noise cuts AI generalisation error
Stronger privacy cuts AI generalisation error in high-noise regimes, yet the robustness-privacy tension returns when noise is low, according to new arXiv prepri
Technology & AI
ALER-TI fills time series gaps using historical retrieval
ALER-TI, a new arXiv preprint, retrieves cached historical patterns to reconstruct missing time series values, tested across six real-world datasets under varyi
Technology & AI
CNeVA traffic sim adds controllability top models lack
CNeVA lets developers steer virtual driver aggression and caution on the Waymo benchmark, offering per-channel control that higher-ranked imitation models lack.
Technology & AI
ActionCache speeds robot AI 34× without retraining
ActionCache cuts inference latency for vision-language-action robot models by up to 34× without retraining, a potential breakthrough for real-time robotic deplo
Technology & AI
G-RRM neural solver hits 33× speedup on Sudoku — with a catch
G-RRM uses neural reasoning models to guide classical SAT solvers, cutting Sudoku backtracking 33× — but the speedup vanishes if the solver can't reject bad hin