[The year is 2031. Unscoped AI development has been frozen at 2029 levels by an international treaty. The growing awareness of the abstract general problem of the abject state of group rationality spurs the creation of a new kind of forum—by, of, and for forum powerusers—aimed at hot-swapping collective sensemaking structures until a good one is developed.] 0722 on the West Coast, I log in. First, I check updates on my posts. Most of them don't warrant interaction from me. Diya Agarwal (real name, verified, like almost everyone on the site) critiqued a critique of a critique of a paragraph in my post. It's a solid critique. I complete-agree-vote the whole critique, and I weak-signal-boost-vote the sentence that contains the main minor insight to anyone who viewed the relevant paragraph in the OP. This only takes a few keystrokes and a few quick glances, because this was one of the top actions suggested by the Lilim that guessed what reactions I might want to give (and the UI...

Loved ones who are gone leave many letters—all the letters of their name, their sayings, the whole book of their life. The letters start to go too. They leave spaces. You start to call her on the phone and remember that she can't pick up, and you can't hear her customary greeting; you think of what you wanted to work with him to build, but that plan is a car with half its parts missing, and tubes and wires sticking out everywhere; the principles you learned from their judgements are impressions left in outline—hopefully enough, but not the same thing. Letters are gone. They leave outlines. At least at first, the outlines say what the letter was, even though the outlines don't have the weight and volume of the letters themselves. Some of the anchors—dates, times, facts—might start to slip or get lost. We can try to put them back, but it's a struggle. Some things aren't so easy to fix to the crypt front, like spirits and principles and dreams and feelings—ways...

What happens when you work closely with someone on a really difficult project—and then they seem to just fuck it up? This is a post about two Chess variants; one very special emotion; and how life is kinda like Chess Bughouse. Let's goooooo! 1. Crazyhouse My favorite time-waster is Crazyhouse Chess. Crazyhouse Chess is mostly like regular Chess. In regular Chess, players take turns making a move, Bishops go diagonally and Rooks go straight, and you try to trap your opponent's King to win the game: (From Lev Milman vs. Joseph Fang courtesy of https://www.chess.com/article/view/10-most-beautiful-checkmates .) In Chess, if you take a piece, it just leaves the board. In Crazyhouse, the difference is that when you take an opponent's piece, you get to use it. Say you take a Black Bishop; then you get a White Bishop in your hand. When it's your turn, you can either do a regular boring Chess move (with one of your pieces already on the board)—or you can drop a piece ...

This some quickly-written, better-than-nothing advice for people who want to make progress on the hard problems of technical AGI alignment. 1. Background assumptions 2. Dealing with deference 3. Sacrifices 4. True doubt 5. Iterative babble and prune 6. Learning to think 7. Grappling with the size of minds 8. Zooming 9. Generalize a lot 10. Notes to mentors 11. Object level stuff 1. Background assumptions The following advice will assume that you're aiming to help solve the core, important technical problem of desigining AGI that does stuff humans would want it to do. This excludes everything that isn't about minds and designing minds and so on; so, excluding governance, recruiting, anything social, fieldbuilding, fundraising, whatever. (Not saying those are unimportant; just, this guide is not about that.) I don't especially think you should try to do that. It's very hard, and it's more important that AGI capabilities research gets stopped. I thi...

[Caveat lector: this is a very long, rambling meditation on concepts and coordination. It's not cut down for size or well-organized. That said, I had several insights while writing it.] 1. Case 1: Is an embryo a person? 1.1. A big blob of subquestions 1.2. Difficult questions produce uncertain, sticky, varied opinions 1.3. A multi-question blob does not have "an answer" 1.4. Coordination about X is fragile to not knowing what other people will think of X 1.5. It's hard to coordinate about big question blobs 1.6. Correction: It's hard to coordinate distributed judgements about big question blobs 1.7. So, to coordinate, people look for simple questions 2. Case 2: Should people be allowed to think freely? 2.1. The First Amendment and delegation 2.2. Erosion of free thought 2.3. Power vacuum 2.4. Simple concepts help coordination because they are anti-invidious 3. Case 3: How nice should you be? 3.1. Nice and kind 3.2. Burning goodness 3.3. The vi...