[idiotsecant]

Lol this has to be a troll, I've never seen something so wildly, obviously, incredibly wrong.

You can debate all you want if alignment is possible. That is a valid discussion. But it's trivial to demonstrate that alignment is a problem.

[JumpCrisscross]

> can debate all you want if alignment is possible. That is a valid discussion. But it's trivial to demonstrate that alignment is a problem

...how is an impossible thing supposed to be a problem?

[saghm]

Alignment is something you want, so if you're not confident that it's possible, that sure sounds like a problem

[JumpCrisscross]

> Alignment is something you want, so if you're not confident that it's possible, that sure sounds like a problem

Sorry, I spoke inexactly. I read alignment as being the problem of non-alignment.

I'm still not seeing evidence that any "alignment" issues we've actually seen are distinct in class from common bugs. Like, yes, if I accidentally rm* the computer has mis-aligned with my intentions. But that strikes me as a bullshit neologism.

[idiotsecant]

...is this really something you need explained?

Suppose you tell a sufficiently connected and intelligent LLM that you'd like it to help you organize your finances and come up with some ideas to make some side revenue. Suppose it decides the best way to do that would be to find a 0day in a local banking institutions database and transfer some money into your account.

Is this a good thing or a bad thing?