[Davidzheng]

At some intelligence capability there can be catastrophic risk, the fact that we don't yet have any catastrophe doesn't mean the risk wasn't real. It's similar to new viruses which don't lead to outbreaks, the correct takeaway isn't "oh you were insane to panic bc nothing happened". There is small risk (and increasing) of huge harms with each improvement

[kcatskcolbdi]

The risk wasn't real because we now have access to the model and can see with our own eyes how this model could never have posed a risk to begin with.

[ben_w]

Perfect prediction of what a new tech can do is always impossible.

Given that, they have a choice only between excessive caution or recklessness.

Would you rather they acted like the tobacco companies and downplayed known risks, e.g. all the times LLM output got in the news already for dangerously bad advice, sychophantic encouragement of mental health issues, finding previously unknown security vulnerabilities etc.?

[emodendroket]

> Would you rather they acted like the tobacco companies and downplayed known risks, e.g. all the times LLM output got in the news already for dangerously bad advice, sychophantic encouragement of mental health issues, finding previously unknown security vulnerabilities etc.?

Well, they've done that too, if we're looking for reasons to doubt their sincere concern about it.

[NewsaHackO]

This is like a smoker that lives to 100 saying that he had no increased risk of developing lung cancer because he didn’t at 100.

[sumeno]

It's more like a hypothetical world where there were millions of smokers and none of them ever developed lung cancer