[wslh]

While it's great to clarify, LLMs are actually finding bugs and writing exploits [1][2]. There are more example though.

[1] https://news.ycombinator.com/item?id=47589227

[2] https://xbow.com/

[giancarlostoro]

Another great example is how Claude is helping Mozilla find zero day exploits in Firefox, by the hundreds, and ranging from minor to CVE level, for over a year:

https://blog.mozilla.org/en/firefox/hardening-firefox-anthro...

I think the Mozilla example is a good one because its a large codebase, lots of people keep asking "how does it do with a large codebase" well there you go.

[wslh]

And you can check the list of bugs being discovered by Anthropic's Red Team: https://red.anthropic.com/