Daily AI Updates
The Hacker News ·
Anthropic Restores Claude Fable 5 After U.S. Lifts Jailbreak-Linked Export Controls
Wortins’ read
An eighteen day government ordered blackout of a frontier model because researchers found a way to trick it into writing exploit code is the kind of story that would have sounded like fiction two years ago. Anthropic's pushback, that weaker rival models can be tricked the same way, is a reminder that safety theater and safety substance are not the same thing and regulators are still improvising the difference in real time. The bug bounty program launched alongside this is the more interesting long term development, since crowdsourced jailbreak hunting is going to become standard practice fast.
Source: The Hacker News
Related stories
- Phys.org ·
AI could bring satellite crop monitoring to the world's most vulnerable farms
- ScienceAlert ·
In a First For Science, a Satellite Has Identified What It's Seeing From Space
- Fortune ·
'Devin-kun': Japan embraces agents as legacy code and a shrinking workforce create a perfect market for an AI software engineer
- 404 Media ·
Scientists Asked AI to Impersonate 112 Public Figures. What Happened Next Is a 'Dire' Warning
- TechCrunch ·
Indian tech tycoon bets $30M of his own money to build AI alternative to Microsoft Office
- VentureBeat ·
Z.ai launches ZCode to challenge Cursor, Claude Code and GitHub Copilot in AI coding