4 ms·
But you're not a member of Anthropic's Red Team, with access to a specialist version of Claude.
by delaminator 6mo ago
But you're not a member of Anthropic's Red Team, with access to a specialist version of Claude.
- staticassertion 6mo agoI don't think that matters at all.
- delaminator 6mo agoI think that Anthropic's own version of Claude will give them different results than the ones you get. "Find zero-day exploits in this popular software." I haven't tried it I suspect that the guardrails will make a difference.
- staticassertion 6mo agoI don't think so. I've never had Claude reject the idea of finding a vulnerability (unlike ChatGPT). The issue is that it's limited by its training set. It'll be trained on things like UAF, it won't be trained on things like "the way your secrets are injected + the way you make HTTP requests + the way you deploy means that an SSRF can expose your private key" or whatever, and that's a technology limitation.