Something something Mozilla CEO who wants Firefox to become an AI product praising an AI model seems fishy. And it's even worse when he acknowledges those 271 vulnerabilities could've been found without the model. Oh well.

That’s a big number but as always the impact of the vulnerabilities is what would be interesting.

Either way that’s a plus for defense.

Or maybe Firefox "code quality" was evident _years_ ago and thats why companies was choosing Chrome codebase ?

Maybe they should try qmail code - bugs or none - that would be something interesting.

Or maybe try OpenSSL. Or Sendmail or Bind :) And what about Exchange ? ;)

If any other project/model had the same amount of money poured into it with this exact same goal, they'd have found it as well. Unsurprisingly, Mythos is effective at executing tasks which it was designed and trained to execute. Turns out, the secret is capital for investment and targeted use-cases - who knew it!

Earlier: https://news.ycombinator.com/item?id=47853649

> Elite security researchers find bugs that fuzzers can’t largely by reasoning through the source code. This is effective, but time-consuming and bottlenecked on scarce human expertise. Computers were completely incapable of doing this a few months ago, and now they excel at it. We have many years of experience picking apart the work of the world’s best security researchers, and Mythos Preview is every bit as capable. So far we’ve found no category or complexity of vulnerability that humans can find that this model can’t.