natural language fluent teddy bears that incidentally scold the child about the One China policy is like... it makes sense given everything else what happened but like... send this to lesswrong in 2015, jesus christ. what have they done to us
12.12.2025 01:51 — 👍 4 🔁 1 💬 1 📌 0
this is not a useful way of thinking about modern methods. most of this space is irrelevant because it's not accessible to us, and we are not uniformly sampling from the subset we can reach
09.12.2025 19:26 — 👍 2 🔁 0 💬 0 📌 0
relatable tbh, I also sometimes feel the temptation to define my problems as impossibly high dimensional spaces
09.12.2025 02:26 — 👍 4 🔁 1 💬 1 📌 0
lots of nostalgia for this image but it was also one of the worst mental models to come out of lesswrong
09.12.2025 01:15 — 👍 15 🔁 1 💬 4 📌 2
stack more layers, got it
09.12.2025 01:11 — 👍 2 🔁 0 💬 0 📌 0
same! but it came out of SSC discourse norms which were already too far gone
09.12.2025 00:39 — 👍 10 🔁 0 💬 0 📌 0
whoops - "ideological turing test", basically "can you write the argument for a given position such that a supporter of that position would think was written by one of their own"
09.12.2025 00:34 — 👍 16 🔁 1 💬 1 📌 1
oh, the joke is that you can choose to capitalize the loan instead of paying interest
09.12.2025 00:20 — 👍 1 🔁 0 💬 0 📌 0
I actually kinda like the intellectual turing test as an exercise, but it was too little and too late
09.12.2025 00:17 — 👍 14 🔁 0 💬 1 📌 0
imo the strongest critique of rationalist discourse norms is that they failed operationalize rationality as a team sport. emphasis on internal steelmanning over structured public debate (and dismissing all related prior art as "rhetoric") is a notable example
09.12.2025 00:16 — 👍 49 🔁 7 💬 1 📌 1
that thread is much, much better than the book it's discussing. (there's some good writing, but misunderstands some of the core arguments, including the titular basilisk)
09.12.2025 00:10 — 👍 5 🔁 0 💬 0 📌 0
🤷♀️ "works on my claude"
07.12.2025 11:39 — 👍 3 🔁 0 💬 0 📌 0
no this is great I'm glad you reposted because I missed it live
03.12.2025 14:22 — 👍 1 🔁 0 💬 1 📌 0
I kinda want to blame TechCrunch for this, all press looks adversarial after a decade of free hypemongering
24.11.2025 02:22 — 👍 1 🔁 0 💬 1 📌 0
aha here is the reaction I was trying to find lol
24.11.2025 02:21 — 👍 2 🔁 0 💬 2 📌 0
that's my point yes
24.11.2025 02:20 — 👍 1 🔁 0 💬 1 📌 0
I have no love for Metz but tech people were hilariously overtuned, misreading very mundane stuff as hostile
My favorite (even sillier) example was tech people objecting to this paragraph from an otherwise glowing article on the xz CVE
24.11.2025 02:17 — 👍 3 🔁 0 💬 1 📌 0
uh oh I've acclimated to Coaticook ice cream and now American ice cream tastes like sugar water
22.11.2025 21:19 — 👍 0 🔁 0 💬 0 📌 0
in the other video I posted he describes M3GAN -> Horizon Line -> The Perfection as "book moves" which is very funny to me
20.11.2025 11:58 — 👍 2 🔁 0 💬 1 📌 0
type of guy who is yimby but only for datacenters
20.11.2025 02:06 — 👍 7 🔁 1 💬 0 📌 0
YouTube video by The Library of Letourneau
The God of Movie Trivia Has Been Unleashed
my wife is more into nl than I am, she recommended
www.youtube.com/watch?v=6S5f...
20.11.2025 00:40 — 👍 1 🔁 0 💬 2 📌 0
putting "get it twisted" at the beginning of this is devious because it's hard to resist mentally hearing rest in nl's voice
20.11.2025 00:07 — 👍 2 🔁 0 💬 1 📌 0
not a good heuristic, but an amusing one: "if I had read this argument before LLMs, what I would I think of it now"
18.11.2025 02:48 — 👍 9 🔁 0 💬 1 📌 0
@norvid-studies.bsky.social claims
15.11.2025 09:50 — 👍 2 🔁 0 💬 1 📌 0
reminds me of
13.11.2025 19:08 — 👍 5 🔁 0 💬 2 📌 0
"for us, the living" is a wild read and people should talk about it more
12.11.2025 00:03 — 👍 1 🔁 0 💬 0 📌 0
(this joke will be funnier in 2-3 years after we have a huge private credit crunch)
09.11.2025 20:40 — 👍 2 🔁 0 💬 0 📌 0
A figment of the universe's imagination. Culture ambassador. http://elsewhereunbound.com
poles of inaccessibility
https://oxidist.github.io
dynomight.net
space invasion
⊙ software eng. FP, type systems, #LeanLang hobbyist. jewish. not a p-zombie i promise.
clarity is a public good
weibac.github.io | 🏳️🌈
Mathematized philosophy of literature
peligrietzer.github.io
temporarily embarrassed orbital weapons platform
Claude says I process my emotions externally.
Thinking about thinking machines | University of Cambridge and Leverhulme Centre for the Future of Intelligence | Previously Google DeepMind
Professor at Wharton, studying AI and its implications for education, entrepreneurship, and work. Author of Co-Intelligence.
Book: https://a.co/d/bC2kSj1
Substack: https://www.oneusefulthing.org/
Web: https://mgmt.wharton.upenn.edu/profile/emollick
Assistant professor (of mathematics) at the University of Toronto. Algebraic geometry, number theory, forever distracted and confused, etc. He/him.
Nathaniel Eliot, first of his name, Jarl of Trollheim, the greater of two evils, the father of punks, the lamenter of code, the schlock and the hard case.
Drawing pictures and working on group coordination. https://danallison.info
Person who does electrical, computer, and music things.
Certified Machine Pervert with compiler-induced psychosis.
Robotanist
they/it/she
📍 london
🦋 emojistats.bsky.sh | 🏷️ https://labeler-starter-kit.bsky.sh | 🟩 https://bluesky-heatmap.fly.dev
🪄 @sortinghat.bsky.sh 🌌 @zodiacsigns.bsky.sh 🧩 @mbti.bsky.sh 🎤 @eras.bsky.sh ✅ @a.rnbw.ca
🤖 @songofsongsbot.bsky.social