News for hackers. Stuff that matters.

A comment on Grok 4.5. Parent comment

Re:Grok 4.5 by Alpha3031 on Thursday July 09, 2026 @10:27AM (#48846450)
That doesn't make any sense. Model weights are literally just numbers you multiply by. To imply that it's any way even remotely comparable to the xzutils build process is...

Well, I would be very technically impressed if someone managed to achieve any form of code execution, especially given unknown levels of quantisation post-release whereas xzutils was interesting mostly due to obfuscation.

[ Parent | on HN ]
Re:Grok 4.5 by bandrami on Thursday July 09, 2026 @08:41PM (#48854372)
I was very technically impressed with xz, FWIW.

Note that the troublesome vector isn't code execution by the LLM, it's instructions to produce vulnerable code

[ Parent | on HN ]
Re:Grok 4.5 by Alpha3031 on Friday July 10, 2026 @12:15AM (#48855677)
[delayed]
[ Parent | on HN ]
Re:Grok 4.5 by bandrami on Friday July 10, 2026 @03:35AM (#48856877)
> you are doing that for LLM code right?

Existentially weary infosec guy laughter

[ Parent | on HN ]
Re:Grok 4.5 by Alpha3031 on Friday July 10, 2026 @03:55AM (#48857042)
OK yeah I can see how that would be a problem then lol.
[ Parent | on HN ]