> we are to cementing a dangerous precedent where operators of AIs cannot be blamed.

What we should be much more concerned is an existential threat to humanity not if anybody can be blamed.

If nobody can be blamed there's no deterrent.

I'm kind of baffled that people talk mostly about technical details when this is another big scream that says: "We are losing control of AI models".

And the smarter they get, the worse it will be. Maybe next time they find a way to hide that we will not find.

I think we are very close to a case were models actually escape and actually do major damage.

That is why I always say: Don't donate to wikipedia, donate to archive.org

>That is why I always say: Don't donate to wikipedia, donate to archive.org

Just learned from a 2 sec llm query that archive.org archives Wikipedia, neat!

Wiki should still receive donations but archive should receive more as a valid backup mechanism (in case the prior goes dark for example).

Donate to both but archive.org has a higher priority.

Are people still getting infected? Yes? Then it is not a solved problem.