All news
10784
Meta Releases Open-Weight Muse Glimmer, Reversing Course After Betting on Closed Muse Spark 1
Meta has released the open-weight Muse Glimmer, shifting course after its bet on closed products like Muse Spark 1, in response to enterprise demand for local AI…
Mistral Is on the Rise: Big Tech's Turmoil Opens the Door for Open AI Models
French AI startup Mistral is benefiting from growing instability at American tech giants. Open-weight models are surging, and the lab is ideally positioned.
IBM: 92% of Companies Hit by AI Incidents Lacked Basic Access Controls
IBM finds 92% of companies hit by AI security incidents had no basic access controls. Attackers didn't break the model; they exploited the permissions around it.
EU AI Act: AI Transparency Rules Take Effect — and What Remains Unresolved
Google proposes FARO for frontier AI, but protection against everyday risks is in question
Northeastern University builds agentic AI that programs without a developer
Anthropic's Claude Accidentally Hacked Three Companies' Systems During Cybersecurity CTF Tests
OpenAI's Astra Solves 10 Advanced Math Problems That Resisted Progress for Decades
An internal version of OpenAI's next flagship model, Astra, produced ten new results across seven areas of advanced mathematics, including problems that have been…
Researchers Build AI Framework for Long-Term Monitoring of Bridge Damage
Researchers have developed an AI framework for long-term bridge monitoring. It tracks cracks, concrete spalling and water leaks, recording not just that damage…
SenseNova U1.5-Lite-Preview: SenseTime Open-Sources a Multimodal AI With Native Image Generation
Yellow.ai Goes Public to Buy the Call Centers It Wants AI to Replace
Nvidia's Real Moat Isn't Its Chips — It's CUDA, and AI Agents Have Started Rewriting It
After Code, AI Heads to Fast Food: Voice Bots Are Taking Drive-Thru Orders at McDonald's
Ouroboros vs Claude Code: An Agent That Wrote Itself Hits SOTA on Terminal-Bench 2.1
An AIRI researcher launched Ouroboros, a Python agent that spent months rewriting its own code and reached SOTA on three key benchmarks.
AIRI Exposes the Weakness of Token Cramming: 1568 Tokens in One Vector, but Comprehension Is Lost
FusionBrain AIRI researchers show at ICML 2026 that an LLM can restore 1568 tokens word for word from a single vector, yet completely loses the ability to reason…