Post B59jcU3RxLd1ywPmbo by itsfoss@mastodon.social
 (DIR) More posts by itsfoss@mastodon.social
 (DIR) Post #B59jcU3RxLd1ywPmbo by itsfoss@mastodon.social
       0 likes, 0 repeats
       
       What do you think of this? Is using AI for fuzzing a good idea?https://itsfoss.com/news/linux-kernel-ai-fuzzing/#linux
       
 (DIR) Post #B59jrLX3vNYdtULw8G by paulk@writing.exchange
       0 likes, 0 repeats
       
       @itsfoss I think predictive ai could be good. That's limited on its task. Still it will always need a human verification.
       
 (DIR) Post #B59kGybbmuLmJjhVKK by enqfila@mastodon.com.br
       0 likes, 0 repeats
       
       @itsfoss I've been reading some papers about it lately. Does have some good results, didn't have time to compare it adequately with other tools (and won't be running experiments to compare them myself)
       
 (DIR) Post #B5AFSMmZKvm344jqXQ by elgregor@librem.one
       0 likes, 0 repeats
       
       @itsfoss Yes. Recently Anthropic found a whole bunch of vulnerabilities in Firefox using AI-assisted methods. https://blog.mozilla.org/en/firefox/hardening-firefox-anthropic-red-team/
       
 (DIR) Post #B5BD1IMdeMFInDX80u by nictakiego@mastodon.com.pl
       0 likes, 0 repeats
       
       @itsfoss No time to wait. Kernel maintainers should start using it, or some bad actors will use it instead - i'm talking about Claude Mythos and future AI hacking softwares. Israel, NSA, Moscow are surely already working on their own Mythos.