Re:Google assigns a CVE for libwebp and gives it a 10.0 score
by toomuchtodo on Tuesday September 26, 2023 @11:06AM
(#37660473)
This is why I'm bullish on LLMs for application security vulnerability searches (like Shodan for code). It'll see what humans don't.
[ Parent
| on HN ]
Re:Google assigns a CVE for libwebp and gives it a 10.0 score
by theptip on Tuesday September 26, 2023 @11:24AM
(#37660714)
Seems to help attackers just as much as defenders though, until you’ve found every single bug (which could be possible, but will likely be expensive.)
Re:Google assigns a CVE for libwebp and gives it a 10.0 score
by malfist on Tuesday September 26, 2023 @11:19AM
(#37660643)
Why would you expect that? LLMs are explicitly trained on what humans do. Humans miss security vulnerabilities all the time.
Re:Google assigns a CVE for libwebp and gives it a 10.0 score
by toomuchtodo on Tuesday September 26, 2023 @11:24AM
(#37660723)
Because the LLM based appsec products I've tested on a large code base have been very good at finding what devs and appsec code reviews didn't.
Re:Google assigns a CVE for libwebp and gives it a 10.0 score
by reitzensteinm on Tuesday September 26, 2023 @11:37AM
(#37660937)
Which ones, out of curiosity?
Re:Google assigns a CVE for libwebp and gives it a 10.0 score
by dnedic on Tuesday September 26, 2023 @12:15PM
(#37661518)
Humans can't be forced to evaluate every piece of code 100000 times in an hour. LLMs can.
Re:Google assigns a CVE for libwebp and gives it a 10.0 score
by philipwhiuk on Tuesday September 26, 2023 @12:22PM
(#37661632)
With infinite hardware, sure. LLMs are pretty expensive to run vs fuzzing.
Re:Google assigns a CVE for libwebp and gives it a 10.0 score
by ZiiS on Tuesday September 26, 2023 @12:17PM
(#37661558)
Humans find security vulnerabilities all the time too. The assumption is a lot of the misses are because highly trained security experts don't have time to carefully read and understand all code without fatigue.