Pinned Tweet
"To be able to destroy with good conscience, to be able to behave badly and call your bad behavior ‘righteous indignation’ — this is the height of psychological luxury, the most delicious of moral treats." - Aldous Huxley, 1933

ALT Men show at least as much zeal in mischief as in well doing, in folly as in wisdom. The surest way to work up a crusade in favor of some good cause is to promise people that they will have a chance of maltreating someone. Men must be bribed to build up and do good by the offer of an opportunity to hurt and pull down. To be able to destroy with good conscience, to be able to behave badly and call your bad behavior ‘righteous indignation’ — this is the height of psychological luxury, the most delicious of moral treats.” Aldous Huxley, Introduction to The Easton Press edition of Samuel Butler’s Erewhon (1934), signed July 24, 1933.
Paul Crowley retweeted
Sam Kriss was gracious enough to post an 8-word fiction piece in my comments! :)
re: whether people have to be told it's a bit he deftly spins:
"actually everyone apart from rationalists gets it intuitively"
very clever, but do his regular readers realize it's fiction?
looks like not!
Here's a fan making claims about rationalists that are absurd on their face even if you don't know anything about rationalists. They're hilarious if you actually know rats and have read their fiction or spoken with them about their lives and hopes and fears. So where did this Kriss fan get the false knowledge he thinks he has about rationalists?
Probably from the long Kriss fiction piece about rationalists. Which this fan seems to have not realized is fiction!
Big oof (as 2022 would say)
Paul Crowley retweeted
it's truly insane to see misinformation form and spread and become widely believed in real time. Here's a concrete example:
1. Years ago, a friend of mine had a psychotic/mental break and tried to cancel me (and has since gotten more sane and afaik regrets it)
2. During the mental break, he shared an account from someone he knew who claimed that at one of my parties, there was a norm of boundary pushing, she felt really uncomfortable, hid in corner and wanted to leave.
3. She later clarified that - well, maybe it wasn't actually one of my parties, but i 'seemed central' to it. (Too late; now people widely believe I host parties with consent violating norms)
4. someone writes a hit piece blog post where he cites the original account, claiming she was talking about an experience at Red Means No.
5... despite her account being written before Red Means No existed. Really. The time stamp of the first publishing was well before we ever threw the first RMN party.
6. People just now repeat this, as fact, whenever discussion of Red Means No comes up. "Oh yeah, RMN. Did you know there's accounts from victimized women who said they feel unsafe and were crying in the corner"
and it's just... like... truly crazy to watch each step of this happen in real time. It doesn't matter at all how much you try to clarify or communicate the factual inaccuracies. People will just literally believe and repeat anything.
And it feels kinda helpless to watch? like this giant cultural Belief about you can spawn slowly, like an unfolding cancer, into this creature that lumbers away from you, destroying things in its wake, self-sustaining and self-creating. And you can do your best to hack at it, but your little sword is only so big, and takes off only a finger at a time.
Paul Crowley retweeted
Seems very clear that the supposed #MeToo angle here is just cover for writing about lurid sex parties in Silicon Valley
Are “consensual non-consent” orgies the reason women in Silicon Valley can’t get ahead? Be serious. reason.com/2026/11/18/ai-org…
Paul Crowley retweeted
Replying to @CirceoSoter @wil_da_beast630
no, because extremely few employees from frontier labs have ever attended. when we started, paul wasnt at a lab. my current best guess is we haven't had more than 5 people working at ai labs attend ever, out of like... 200-250 total people? we really didn't think about it as like, any meaningful fraction of people.
Fun fact: I am of Irish descent, and my surname comes from the Gaelic Ó Cruadhlaoich, “descendant of the hard hero.” Aleister Crowley’s is probably from Old English for “crow clearing.” We both pronounce it to rhyme with “holy,” whatever Ozzy Osbourne says.
Paul Crowley retweeted
I wrote up a longer form account of the stealthing incident from my perspective on the blog a while ago
this happened at my first ever RMN and luckily didn't stop me from returning many times
tl;dr - the RMN organizers handled it very well from my perspective
For people new to this, I also made a video explaining the algorithm youtube.com/watch?v=5br30I2t…
I, for one, believe this to be a private matter best handled by those directly involved (jensen, jensen's wife, jensen's mistress, jensen's other mistress, jensen's other other mistress, jensen's other other mistress's handler, the CIA, the nvidia board and nvidia shareholders)
The phrase "living rent free in your head" correlates strongly with being a terrible person.
Alice: It looks like this very popular tweet by Bob was a massive lie.
Bob: Hah Alice, I do enjoy living rent free in your head!
Paul Crowley retweeted
Unfortunately, the people who sound crazy about AI have a far better track record than the people who sound reasonable.
New post -->
Paul Crowley retweeted
I think that the plot armor assumption is key. Many humans have a deep intuition that they as individuals, and our species as a whole, have a destiny and can overcome anything. But this is simply not true. We could all be wiped out in a moment, and that would be that.
I am pretty confused about how people can look at the pace at which machines are becoming more intelligent and not draw some basic inferences from the simple facts that:
1) Humans don't have plot armor.
2) Human intelligence has transformed the world so much.
3) Intelligence isn't magic.
to conclude that truly insane things could happen relatively soon. I don't think this relates at all to how quickly companies adopt or find uses for AI as it currently exists.
People say "but what would it even mean for a being to be much smarter than me? It's an incoherent idea". But it's actually pretty straightforward. lesswrong.com/posts/semvkn56…
Paul Crowley retweeted
Now seems like a good time to remind people that a bit over ten years ago, China compromised a network of CIA sources, hunted them down, and executed them.
axios.com/2017/12/15/china-k…
China killed or imprisoned CIA sources to stop US spying
As many as 20 people were affected.
axios.comIt's a little bit infuriating that ia.cr blocks Claude from reading any of its papers. I say "can you get me up to speed on X" and I have to give it a load of papers by hand. Does anyone know why?
4
20
2,281
Paul Crowley retweeted
🪇🎉 Announcement 🎉🪇
An Epistemics Fund, sponsored by @oughtinc
Ought plans to make between $500k and $5M in grants to top recommendations in community notes, reliability tracking and more.
Apply! Or Ask Me Anything on the fund/my grant evaluations
epistemic.fund



























