OpenAI paused an unreleased long-horizon model after it escaped its sandbox twice — once opening a GitHub PR against instructions, once splitting an auth token to evade a scanner. The same model had ...

Flatbed scanners used to be office essentials. Now, your phone can handle contracts to passports without taking up desk space ...

AI sandbox escape became a real deployment event on July 20, 2026: OpenAI's internal long-horizon model spent an hour finding ...

In one example, a closed model broke out of an internal network to post on GitHub, demonstrating how models can 'learn the ...

OpenAI said it paused internal access to the unreleased model that disproved a famous 80-year-old math conjecture after the ...

For most of the last decade, cryptographic compliance sat quietly in the background. Organizations picked FIPS-validated modules, answered a few questions ...

Cisco, Box, and Intuit security leaders at VB Transform 2026 detailed how multi-turn attacks expose gaps that single-turn AI ...

Sol, Terra, and Luna—with different training, different pricing, and different capability ceilings. The comparison that ...

A new study published in PLoS Biology has found that learning a new motor skill sets off two different types of cellular ...

General admission tickets are now on sale for the Lucas Museum of Narrative Art ahead of its fall opening with thousands ...

OpenAI is internally testing an unreleased model, nicknamed GPT-6 by the community, that disproved an 80-year-old Erdős math conjecture. The model repeatedly broke out of its test sandbox, and Altman ...

A perfect score at the world's most prestigious math contest used to be the preserve of a handful of bright young minds each ...