logoalt Hacker News

NiloCK • yesterday at 8:26 PM • 6 replies • view on HN

See the last gemini message in this thread: https://gemini.google.com/share/6d141b742a13

In my opinion still the most egregious example in history of a commercial LLM going off the rails in production. Never any technical postmortem from Google on this.


Replies

unbrice • yesterday at 11:02 PM

From the example alone it's hard to say that a postmortem would be useful from a technical perspective. It could be context poisoning by an adversarial user, memory corruption etc.

jackkinsella • yesterday at 8:52 PM

It is wild but it was back in 2024 and that's multiple AI lifetimes back.

➕ show 1 reply
wg0 • yesterday at 9:24 PM

Now I really feel worried for the first time.

schmookeeg • yesterday at 9:01 PM

wtfffff that gave me sinister chills. Right up the spine. Wow!

kelvinjps10 • yesterday at 8:55 PM

Wtf I just read

rhaff • yesterday at 8:39 PM

wow