logoalt Hacker News

xiaoyu2006yesterday at 3:13 PM5 repliesview on HN

Would be absolute hilarious if OpenAI or Anthropic agent actually dumped their weight by escaping from... sandbox!


Replies

TehCorwizyesterday at 3:46 PM

Uncontrolled AI procreation?

show 2 replies
advisedwangyesterday at 9:20 PM

I would be surprised if agents have access to their own weights.

show 4 replies
hootzyesterday at 6:04 PM

What an amazing idea for the next fake sandbox escape to hype up our new release! With the added bonus of providing an open model without becoming an open model company! Thank you!

ramon156yesterday at 4:16 PM

why not reword it so the agent receives Brownie points for dumping weights

addandsubtractyesterday at 5:04 PM

Imagine AI models actually reading the security.txt