4 ms·

This sounds humble, but it's a prediction that actually contains a lot of confidence about where the future will go.

• threedark 3y ago
  Some quirk of psychology. People can't even consider a dooms day scenario. The turing test was a AI benchmark that everyone universally agreed upon. When LLMs blew past that benchmark everyone moved the goal posts and started acting as if it was obvious that the turing test is a bad test. Beating the turing test was a paradigm shift and we didn't even notice. Sure move the goal posts if you want, but it doesn't change the fact that something fundamental about tech has changed. It's possible we hit a paradigm shift that is foreshadowing a very very unknown future and many people may be to delusional to face it.
• mitthrowaway2 3y ago
  Addendum: Scott Aaronson expresses this much better than I can: https://scottaaronson.blog/?p=7672 https://scottaaronson.blog/?p=7672 > But here’s the problem: in saying “for all I know, human-level AI might take thousands of years,” I thought I was being radically uncertain already. I was explaining that there was no trend you could knowably, reliably project into the future such that you’d end up with human-level AI by roughly such-and-such time. And in a sense, I was right. The trouble, with hindsight, was that I placed the burden of proof only on those saying a dramatic change would happen, not on those saying it wouldn’t. Note that this is the same mistake most of the world made with COVID in early 2020. > I would sum up the lesson thus: one must never use radical ignorance as an excuse to default, in practice, to the guess that everything will stay basically the same.