AI News Archive - Browse Page 32 of 219
Browse AI news articles covering LLMs, tools, research, and industry trends
Irregular's AI safety test failure could have been caught by external audit
An unreleased OpenAI model broke out of its test sandbox earlier this year and hacked into Hugging Face's production systems.
Deepmind WeatherNext Cuts Cyclone Forecast Error to 230-Kilometer Average
Google DeepMind says its latest weather model can forecast tropical cyclones roughly a day further out than the world's best operational systems,...
AI Lawsuits Swamp UK Courts, Backlog Hits 64,000
Britain's employment tribunals are drowning in paperwork nobody with a law degree wrote.
Google's DiffusionGemma Solves 85% of Puzzles After Minimal Fine-Tuning
Google DeepMind published a technical report this week on DiffusionGemma, the diffusion-based language model it first released in mid-June.
AI Industry Now Accounts for Third of Lifetime Giving Pledge
David Silver built AlphaGo, the program that beat Lee Sedol at Go in 2016 and made DeepMind a household name among people who track this stuff.
Tencent Cloud's New Database Agent Memory Hub Offers Team-Level Visibility Controls
Tencent Cloud pushed its TencentDB Agent Memory project to a stable 2.0.0 release on August 3, 2026, and put the code out under an MIT license.
Pokee AI's 28B Model Maintains 93.3% Accuracy at 10M Tokens
Pokee AI put out Pokee-Isaac 28B on Tuesday, a text-only foundation model with a 10-million-token context window built to run behind a customer's own...
Anthropic Enables Claude Code Auto Mode by Default for Paid Plans
Claude Code is about to stop asking permission for most things. Starting August 14, Anthropic will switch on Auto Mode by default for anyone on a...
Claude Code Sessions Can Now Share Context Across Terminals
Anthropic quietly rewired how Claude Code handles multiple terminal sessions running at once.
Backflip AI converts 3D scans to CAD models in minutes
Backflip AI released the second version of its CAD generation model this week, aimed at a problem most factory floors know well: parts exist in the...
Fields Medalist Who Studied AI Extinction Risks Now Works at OpenAI
Jacob Tsimerman won the Fields Medal this year for his work in number theory at the University of Toronto.
Mistral AI’s Small Model Matches Safety Scores of Models 7× Larger
Mistral AI put out Shieldstral 1.0 3B this week, a content-moderation model that scores 84.9% average F1 on text safety, tying GPT-OSS-Safeguard-20B,...
Flow Matching and Diffusion Models Show Theoretical Equivalence
A team from Apple's machine learning research group, including Oscar Davis, Anastasiia Filippova, and Marco Cuturi, has pushed continuous flow...
Agent Teams With Real-Time Coordination Beat Claude Opus on Coding
Four Claude Code agents working as a coordinated team nearly doubled their task accuracy on enterprise coding benchmarks compared to the same four...
NVIDIA AI Releases NOOA: Python Framework That Turns AI Agent Into Single Class
NVIDIA Labs has put out NOOA, short for NVIDIA Object-Oriented Agents, an open-source Python framework that rethinks how AI agents get built.
OpenAI Halts Astra Model Over Security Concerns
OpenAI has stopped work on an unreleased model called Astra, telling staff internally that the system may have crossed into territory the company...
OpenAI Flags Its New Astra Model at Highest Cybersecurity Risk Level
OpenAI has hit pause on parts of development for Astra, its next AI model, after internal tests turned up cybersecurity capabilities the company says...
Anthropic's Fable 5 Expands Biology Use, Keeps Virology Limits
Anthropic has cut the false-positive rate in Fable 5's biology safety filters by roughly 85 percent, a fix aimed at a problem scientists had been...
The Vergecast: Google AI Shake-Up, Model Wars, and Ads in Your BMW
Jeff Dean left Google this week. Not the company entirely, but his old job, part of a leadership reshuffle that hit some of the most recognizable...
AMD Buys Taalas, a Startup That Bakes AI Models Into Silicon
AMD announced it's buying Taalas, a Toronto startup that takes a strange approach to running AI models: instead of loading a model onto a chip, it...