Worse, P(doom|corrigible human replacement level AGI) is also very high, because the dominant strategy in international competition will involve zero humans. So we are screwed even if "alignment" turns out to be surprisingly easy or if there is a plateau right around human level.
The most likely path I see for human survival and flourishing is for AI to do something in the near future so destructive and terrifying that it unites humanity behind permanently banning it at any cost.
It therefore seems like "AI safety" people may actually be working against our safety.
The first thing I do when I see an interesting article title is click on the comments to see if people have noticed that it's AI generated. If I didn't have this option because of the policy you desire, I would probably just stop reading HN entirely.
How about i^4x? It makes it clearer which direction we are rotating (vs -1 which is a 0.5 turn rotation in either direction) and avoids the garden path confusion that could arise from it not being obvious from the left side of the equation that we are working in the complex plane at all.
> AI writes copy like this because that's how copywriters did it in the training set
I don't think this is true. Someone probably had to write like this, but a base model has to be able to write like anyone. The specific mannerisms of AI appear in post training, either as things that are rewarded or as mode collapse artifacts. ChatGPT didn't get obsessed with goblins because that's incredibly common among humans.
Right. But because it’s the smallest structure of its type (speaking loosely) it feels like something we should have a grasp on, even though it contains more complexity than we could ever describe or compute with (since both of those are countable.)
