Open Thread 208

...

This is the weekly visible open thread. Odd-numbered open threads will be no-politics, even-numbered threads will be politics-allowed. This one is even-numbered, so go wild - or post about whatever else you want. Also:

1: Comment of the week is from Richard Ngo, who helpfully corrects some of my discussion of his dialogue with Eliezer Yudkowsky:

1. Scott describes my position as similar to Eric Drexler's CAIS framework. But Drexler's main focus is modularity, which he claims leads to composite systems that aren't dangerously agentic. Whereas I instead expect unified non-modular AGIs; for more, see https://www.alignmentforum.org/posts/HvNAmkXPTSoA4dvzv/comments-on-cais

2. Scott describes non-agentic AI as one which "doesn't realize the universe exists, or something to that effect? It just likes connecting premises to conclusions." A framing I prefer: non-agentic AI (or, synonymously, non-goal-directed) as AI that's very good at pattern-matching, but lacks a well-developed motivational system.

[continue reading full comment here]

2: Some people have requested guidance for when you can advertise your own blog/website/etc in the comments here. I would say: on regular posts, only if it’s something very relevant, so relevant you would post it even if it wasn’t yours. On open threads, try to limit yourself to twice a year, and only if you think it will spark a genuinely interesting discussion. On classified threads (which I want to remember to do more often), go wild.

3: Don’t forget to resubmit and summarize your proposals for Grants ++, if that is a thing you want to do.