Pinned Tweet
The moral panic about AI mirrors the opposition to “Penny Dreadfuls.” They believed this literary slop (“pulp fiction”) would degrade society. Instead, the lasting legacy was mass literacy and the creation of pop culture.
6
4
72
5,044
this kind of talk is in such bad faith. "oh, it's hard to understand the proof, unlike past breakthroughs that were immediately written perfectly and disseminated to the world within days without any work." not that NPR minds bad faith. worthless enemies of the truth
Fascinating article about Navier-Stokes solution and what mathematicians have learned … “so far humanity has learned very little from the solution. Although they believe the proof to be technically correct, the dense, 166-page manuscript drafted by AI is proving to be a difficult read. "So far it's been very difficult to really extract any human understanding from this new AI proof," said James Maynard, a mathematician at the University of Oxford. The paper is not written for humans," said Javier Gómez-Serrano, a mathematician at Brown University who uses AI in his own research. He said he believes the proof could help advance the field after "some serious re-writing," but "as of today, the paper doesn't teach us much."”
npr.org/2026/09/22/nx-s1-596…
2
92
Yann LeCun: we are not going to get human-level intelligence or AGI by just scaling LLMs.
“There’s absolutely no way in hell.”
“The idea that we’re gonna have a country of geniuses in a data center, that’s complete BS.”
“It may feel like you have a PhD sitting next to you.”
“But it’s not a PhD.”
“It’s a system with gigantic memory and retrieval ability, not one that can invent solutions to new problems.”
1
377
obviously a US-China teamup would be completely unstoppable. who are we gonna ice? i'm thinking Mongolia. i'm not sure if China ever got fully even for Genghis Khan, and i resent the price of cashmere
🚨 UPDATE: Chinese President Xi says he fully believes Thucydides trap — inevitable WAR between the US and China — can be avoided
"Our two militaries should maintain regular dialogue and improve mechanisms for crisis communication and prevention. The Thucydides trap can be overcome."
"In the second world war, Chinese and Americans fought shoulder-to-shoulder. Chinese people have never forgotten that part of history, nor have the American people."
Xi has been issuing remarks of unity with Trump and Melania Trump sitting behind him
Trump and Xi will now continue the state visit, with a DRILL PLATOON REVIEW happening next
6
443
shaun retweeted
A style guide should not require ideological commitments that require one to believe that solving Millennial Prize Problems can be done without thought or understanding.
2
8
208
suppose there was a button, and if you pushed it, an AI lab would produce an important math result, but also a random anti-AI mathematician would suddenly explode into a rain of gore. would you push it? for collatz, goldbach, twin primes, odd perfect numbers, p vs np, heck pretty much any complexity separation, push, push, push? you'd be spamming that button until you got carpal tunnel syndrome, wouldn't you? you are one sick puppy. me, i would never do that. nope. wouldn't even think about it. no sir.
2
13
616
This guy masterfully expresses the nature of the “ick” I get from Claude. However, I’ve seen a couple examples suggesting Opus 5.5 is improved. Anyone still getting hit with pretentious windbaggery?
I absolutely hate the way Claude (Fable and Opus) writes - it’s so full of itself, so fake-eloquent without saying anything, so know-it-all, presenting the simplest insights in a repulsively smug tone - and it seems impossible to get rid of all these Claude-isms it produces even with strict global instructions.
Can’t stand any more “load-bearing assumptions”, any sentences that end with “and that’s the one point that lands with me”, generally any short sentences that start with “And” and a complete over-use of references to “the one XYZ”.
It feels like AI for people who have never learned how to communicate succinctly, and who constantly have the need to feel smarter than everyone around them (hint: they’re not smarter, they’re just insecure).
ChatGPT, Grok and Kimi are miles ahead on producing a more natural communication style. The actual analysis, the automations and the artifacts might (often) be better - but Claude just feels so repulsively unlikeable to interact with.
1
1
273
word on the street is that opus 5.5 has made a measurable step away from giving the ick. looking forward to seeing the scope of the improvements
"the whole thing rests on one idea"
"three things define it"
"there are two flavors"
"want a floor under them"
"you've made $1,200 on a $300 outlay"
"that asymmetry is the shape worth internalizing"
Pure uncut claudeslop
1
141
i mean he just didn't think LLMs would work, and thus thought such calls were stupid. it's like people who are frustrated with environmental regulations because they think it's a corrupt grift versus people frustrated with environmental regulations because they are straight-up climate change deniers.
2
395
shaun retweeted
At the current rate of AI progress I predict that it'll be 2027 in just a few months.
93
234
4,455
125,118
let me translate this for the confused shrimp-counting "rationalists" who don't understand realpolitik: "if you do dumb shit that fuels the ai doomer narrative democrats are now running political campaigns on, we'll fuck your shit up. the marketing stunt is not worth it, and you will not get regulatory capture. you've been warned."
🚨 TREASURY SECRETARY BESSENT JUST NUKED ANTHROPIC & OPENAI’S “ROGUE AI” IMMUNITY SCAM ON LIVE TV
"A sitting employee came out, said there's a 10% chance of an extinction level event. But then the labs also said, take the liability off of our hands. And we will NOT do that.
The Hugging Face incident is the responsibility of OpenAI management, NOT a bunch of agents.
It is humans who are responsible, not the AI.
These labs need to take responsibility for themselves. They can slow down any time they want to."
CHECKED.
1
4
24
969
We’re working with an independent advisory group of mathematicians to help OpenAI responsibly share advances in AI and mathematics.
The group will advise on how we assess and communicate new mathematical results, uphold academic and professional standards, and build tools that support mathematical research and learning.
Through this work, we want mathematicians to be at the center of shaping how AI supports mathematical understanding and how its benefits reach the wider community.
openai.com/index/advisory-gr…
9
16
222
13,159
