Thanks for sharing this.

Dear people who read this and agreement-downvoted (ETA: wrote this cause above comment was well in the agreement-negatives at the time of writing): Do you think this isn't Cremieux's account, or that the quoted example is an acceptable thing to say, or what?

Eh, he was there last year, I figure he might well go again. If I happen to hear that he's definitively not attending this year (or, idk, if he ends up attending as a regular guest and not an Invited Author Guest, I take less umbrage with that) I'd love to go. 

I was honored to be invited again to this year's LessOnline - I really enjoyed the last one. However, I'm going to turn down this invitation as I'm uncomfortable being in the same company of invited author guests as Cremieux.

I didn't know who he was last year, so after hearing concerning murmurs from various places, I looked into his work. Hoo boy. I don't think that being interested in genetic differences between ethnic groups necessarily makes one racist, but I think it's the kind of area where you have to be extraordinarily careful to proceed with caution and compassion and not fall into racist fallacies (coexisting in a terrible cycle with shoddy scholarship). I do not think Cremieux meets this standard of care and compassion. 

Also, I get the sense he's generally a jerk to those around him, which is not as big of a deal but is not helping. He reacts to challenges or criticism with insults, over-the-top defensiveness, and vitriol. 

I don't like what he's about, I think the rationalist community can do better, and I do not want to be a special guest at the same event he's a special guest at.

I hope that LessOnline goes well and that those who do go have a great time, and that my assessment is completely off-base. I mean, I don't think it is, but I hope so.

Delightful! I DO enjoy knowing that!

Advice: The AI-generated diagram here doesn't add anything and in fact indicates strongly that I wouldn't want to read the post. One of the things about diagrams being so important and eye-catching associated with writing is that they communicate information, so if a diagram is clearly half-assed and wrong, it makes one assume that the text is too. (Half-assed is maybe not the word - minimally-assed? MS Paint stick figures would be fine here, for instance.)

There's extraneous detail. The text is garbled and irrelevant.

 I think if you use image-generating AI to make diagrams you should then edit it afterwards to make sure it's actually, like, good and represents what you wanted, and add your own captions.

That's definitely a good point and model vis-a-vis "this group/ideology is targeting these people specifically".

I would also point out that specifically rejecting demographically-vulnerable people is likely to push more of them towards this ideology - though even if that effect weren't in play, it would still be shitty to tarnish a broad group of generally fine community members by common demographic.

I think this is a horrible thing to say. The murderers are associated with each other; that gives you much more information than just knowing that someone is trans or not. There are many, many stellar trans rationalists. I'm thinking you maybe are thinking of the standout dramatic cases you've heard of and don't know a lot of trans people to provide a baseline. 

I don't disagree with you about not wanting to read LLM output, but:

> Everyone in Cyborgism or AI Twitter or LW who talks a lot about talking a lot to LLMs for generic conversation, rather than specific tasks, seems to lose their edge and ability to think critically

- is a very strong claim to just throw out there. Everyone? Are you sure you're not remembering the people who stand out and confirm your theory? You're getting that they're (for twitter users) "losing their edge and ability to think critically" from, like, tweets?

I'd suggest writing about stuff you're interested in but that don't feel crucial to get right, if that makes sense. A hobby, fiction, stories from your life, about your day, funny observations...

If you don't have any other interests and just have to write about unimportant boring stuff - hey, yeah, sure, polish turds. I'm reading Ulysses right now and it's, like, mythologizing some guys going around their everyday lives and drinking and being casually rude. And it's one of the most beloved novels ever. Writing about boring everyday bullshit in ways that sound cool is a time-honored tradition. 

Well, okay, you can also start writing about things you really care about - but I feel like there's a kind of person who might read this who, like, has a thing they really care about - "we need to develop more mRNA vaccines", maybe - and is going to write a mid essay about mRNA vaccines, and then they'll sadly think "well, nobody liked that essay," and never go back to it - and that would be sad. So if you're going to practice via writing things that are very important to you, you might have to be willing to write on the same topic/thesis a few times. 

(Also, if a person in your audience reads one essay from you and doesn't like it, they might not be willing to read a second essay from you on the same topic even if it's better now - so you might also want to show different iterations to different audiences, if your potential audience isn't large. YMMV.)

Yeah, so I bet passive osmosis has in fact gotten you somewhere, but to go a bit beyond that -

• Can you identify when you're reading writing you like vs. writing you don't like?
• What's the difference?
• What kind of properties does writing you like have, compared to other writing? (Especially compared to writing that's "just okay", as opposed to actively bad)
  • Can you recreate these in your own writing?
• What effect does good writing have on you? (This is sort of an art more than a science, but like - do you understand the thing better? Do certain sentences just like really hit you? What's going on there?)