An autonomous agent took 17,600 actions to breach Hugging Face's infra over 4.5 days.
When it came time to investigate, the guardrailed frontier models refused to help — their own safety training read forensics as an attack. An open model did the job instead.
Inspect the model. Exploit-test it. Certify it. Contain the agent.
49 names for this exact category — launching the board:
https://t.co/POFPaFKKsK
🚨 NOW: Anthropic chief DARIO just walked up and said President Trump helped put his Super Intelligence worries at ease
"If we do this right, we work with the President and everyone here, we can win safely!"
Imagine that, all it took was DONALD J. TRUMP getting involved and the tables are turning!
No more SI dooming, no more surrendering to China.
America will remain number one BOTH for national and economic security, as long as Trump is president 👏🏻
@namefi_io AI Model Assurance.
https://t.co/oQl7FSPbad
https://t.co/aeNQ7lUxSx
https://t.co/FcACtz4Hrs
https://t.co/OLZYtDreu9
https://t.co/RNe2HMqovD
Several domain clusters can be found here:
https://t.co/TtVthx3A5K