Published on February 5, 2025 4:30 AM GMT
Yeah I put this off until the last day, and I'm not sure this is the format Raemon was actually looking for. Oh well.
Then, in proportion to how valuable they seem, spend at least some time this month reflecting...
...on the big picture of what intellectual progress seems important to you. Do it whatever way is most valuable to you. But, do it publicly, this month, such that it helps encourage other people to do so as well. And ideally, do it with some degree of "looking back" – either of your own past work and how your views have changed, or how the overall intellectual landscape has changed....on how you wish incentives were different on LessWrong. Write up your thoughts on this post. (I suggest including both "what the impossible ideal" would be, as well as some practical ideas for how to improve them on current margins)...on how the LessWrong and X-risk communities could make some group epistemic progress on the longstanding questions that have been most controversial. (We won't resolve the big questions firmly, and I don't want to just rehash old arguments. But, I believe we can make some chunks of incremental progress each year, and the Review is a good time to do so)
I. On the big picture of what intellectual progress seems important to you
I've said it elsewhere, and I'll say it again here: the thing that hooked me is The Martial Art Of Rationality. I'll do my proper rant about that elsewhere, but that's the direction I'm pointed in. I want progress in the form of clear examples of ways brains make obvious mistakes paired with drills on how to practice not making that mistake, and a way to check if the drill is working.
Anki and spaced repetition, to stop forgetting things.
Calibration training tools, like the introduction of Fatebook.
Even focusing on checking if things are working by focusing on the feedback loop.
I did a lot of meetup writeups for Meetup In A Box. And yet, this probably isn't much progress? It's hammering pitons into the mountainside to make it easier for more people to climb up, not braving a fresh climb somewhere nobody's been before.
Raemon's Feedback Loop Rationality is a central example of what I want more of.
II. On how I wish incentives could be different on LessWrong
I wish there was more reason for great writers to stick around.
Stop for a moment and think of the great rationalist writers. Grab a piece of paper or open a text doc and write down five names. Don't feel like you have to take a long time, this can be like ~30 seconds.
Did you?
I'm going to leave some space, page down when you're ready.
.
..
...
....
.....
......
.......
......
.....
....
...
..
.
I'd go Eliezer Yudkowsky, Scott Alexander, TheUnitOfCaring, Zvi, Duncan Sabien. If I go for another five, I'd say Alicorn, Gwern, Elizabeth Van Nostrand, Sarah Constantin, Jacob Falkovitch. If you came up with a different list, that's fine.
How many of them write on LessWrong?
Scott and Duncan have Substacks. Sarah, Jacob, and Zvi also have Substacks. Alicorn, Gwern, and Elizabeth have their own websites. TheUnitOfCaring is writing elsewhere.
Zvi crossposts to LessWrong. Gwern does, but most of his best writing isn't here. Elizabeth and Sarah Constantin both write here, mostly crossposting. Jacob used to crosspost, but stopped a couple years ago.
As far as I can tell, Eliezer mostly writes on twitter these days. I get it, if you're trying to do public outreach then you go where the public is, but that seems kind of embarrassing from LessWrong's perspective.
If there was a fresh great writer these days, why would they post to LessWrong? If you're trying to keep up with the rationalist writing, I think a Substack account would be more useful to you than a LessWrong account.
Failing a way to incentivize those writers to LessWrong, maybe incentivize curation? Grab the good Eliezer tweets and bring them back to LW, collect half credit for the upvotes maybe?
III. On how the LessWrong community could make group progress
The question Raemon asks includes X-risk as a community, not just LessWrong. I don't have any answers for X-risk. Sorry.
That points in an important direction though. I think "The LessWrong Community" is too big, and pointed in too many directions. It might be more useful to think of us as a dozen smaller communities in a trench coat. The trench coat is important, it keeps us close enough to recognize each other and make finding each other easier.
Towards that end, having common knowledge seems useful. I'm increasingly fond of the EA Handbook. I'm told EA student groups go through the handbook once a year or so, and cohorts of new EAs get introduced via the handbook. This means if you can actually somewhat assume your fellow EAs know the phrase "marginal impact" or a basic understanding of why charities can have different amounts of impact.
Here's something embarrassing. There's a bunch of Yudkowsky's Sequences posts in the EA Handbook. Pick a LessWrong meetup attendee and an EA meetup attendee at random. I would be more confident that the EA attendee has read Making Beliefs Pay Rent than the LessWrong attendee.
I've been thinking of writing up some kind of Five Things Aspiring Rationalists Should Know document and try to get every meetup group to read the thing. I expect there'd be some quibbling over exactly what got added or what a close sixth thing would be. That seems fine. It would be important to keep this document short, since "read the sequences" is a bit much. The Highlights From The Sequences is close to what I want here? The idea would be to be able to assume the people in the room (or commenting on a LessWrong thread) had some common building blocks.
Related: I think it would be worth identifying a handful of rationalist skills, figuring out how to evaluate them, and checking if anything we're doing is helping improve those skills. That's not a step forward into new territory, but I think it'd help catch up.
That doesn't mean we can stop- you will always have the 101 space- but we can't take a step forward together unless most people actually have caught up to the step we're currently on. If there's some objection that everyone points out the first time they hear an idea, but the objection is incorrect, then this lets you get everyone past the objection and on to the second or third problems with an idea instead of just the first. If we can't take a step together, then we should split up into groups small enough that we can get on the same page.
IV. I propose an archipelago
I think friendly, gentle splintering might be in order.
Seven years ago, Raemon floated the idea of LessWrong as a Public Archipelago. I'd like to see more movement in that direction.
My martial artist's view of rationality, focusing on drills and concrete subskills, isn't the consensus view. There's a lot of AI content on LessWrong these days, and I basically tune it all out. (Thank you to the LessWrong team for the filtering tools to do that!) How much AI content is there? Here's a screenshot of the Latest page right now, from an incognito window:
Welcome to LessWrong, AI, fiction about AI, AI, AI, fundraiser, AI, AI, AI, AI, AI, musings about moderation policy, book review, AI, AI, AI. I'm trying to filter out 3/4ths of the site at this point.
Give me a dojo.lesswrong.com, where the people into mental self-improvement can hang out and swap techniques, maybe a meetup.lesswrong.com where I can run better meetups and find out about the best rationalist get-togethers. Let there be an ai.lesswrong.com for the people writing about artificial intelligence. I'm not against a little cross pollination even! Maybe the other subsections can send over one or two posts a month, as a treat.
Let the islands of the archipelago be styled a little differently, or have their own intro and faqs. If someone's a particularly prolific writer, give them their own subdomain. If it's not too hard to figure out, make it easy to pay the local writer. C'mon, acx.lesswrong.com would be pretty great. For that matter, why does it need to be a subdomain? I'd love a Screwtape.com that pointed at a LessWrong page, especially if I could have more freedom with the user page than LessWrong generally allows.
I suspect an instance of LessWrong, configured to be styled how they wanted it, with similar affordances for subscriptions, clearer moderation controls, and making the one writer's posts front and centre, would be able to put up more of a fight against Substack.
(Gwern.com is an exception. That's a piece of web development art. Scott and Sarah and Zvi and Elizabeth though, my guess is the most significant advancement since Wordpress is the normalization of the Subscribe button giving them money.)
This would let smaller groups identify themselves and differentiate more. From what I can tell, it's normal to archive binge a blog's past posts, but not to archive binge all of LessWrong. Even a "next post" button on the bottom of all of, say, Raemon's posts or Habryka's posts would be interesting. Right now to do that, I'd have to go to their user page, click a post, read it, click back, and find the next post. I'd be interested in seeing some writer's circles maybe, a new Inklings writing on the same domain the way that Hanson and Eliezer shared Overcoming Bias for a while.
Maybe this users in a second rationalist diaspora. There is something useful about having a single conversational locus. My observation right now is that LessWrong isn't the place where all the great thinkers and writers are making progress though, and it's not where the readers stay to get everything they want to read. If this worked, you could bring back some elements of the community that never returned. There's fairly little fiction published on LessWrong for instance, despite HPMOR being such a big influence and the health of r/rational.
There's my review and pitch. Hope it's helpful.
Discuss
