"Often with contextual hints" is doing some heavy lifting here, IMO. I agree with the article's premise -- you don't need Mythos to use AI to find novel, complex vulnerabilities -- but these results as presented are somewhat misleading.
"(...) a well-designed scaffold naturally produces this kind of scoped context through its targeting and iterative prompting stages, which is exactly what both AISLE's and Anthropic's systems do."
The current tech is a sigmoid and even using the abilities of the AI, novelty, improvements don't appear to be happening at any exponetial pace.
What makes you say that? I'm only asking because the data I've seen looks pretty cleanly exponential still, e.g. https://metr.org.
baq•2d ago
Notably, Kimi K2 and GPT-OSS-120b do quite well when provided with the isolated context. Article seems to be heavily LLM-assisted, but the content itself is good.