← Back to live feed1 story
Anthropic confirmed that sensitive AI model weights, source code, and user conversations were not accessed during a recent internal network breach, which it said took researchers from Cathedral several days to exploit.
Sign in to suggest
Key sources
- SOURCE@lizzadwoskin“including Slack channels where the most sensitive models are discussed”x.com
- SUPPORT@choblin29“Anthropic confirmed access to a limited set of sales, HR and IT systems, but said Claude conversations, source code and model weights were not accessed”x.com
- SUPPORT@gerritd“Cathedral is made up of ex-DOGE staffers”x.com
- SUPPORT@logangraham“this wasn't an ordinary bug bounty finding – this involved exploiting 3rd parties and accessing data”x.com
- SUPPORT@niubi“the Cathedral report on its access inside Anthropic had circulated within the government and raised national security concerns”x.com
- SUPPORT@zackkorman“Got access via some third party SaaS Anthropic uses”x.com
- SUPPORT@francoischauba1“this story seems buried by the s1 news..”x.com
- SUPPORT@hesamation“Researchers say they hacked into Anthropic’s internal systems in under 12 hours using agents”x.com