
Simon Pepin Lehalleur
@plain_simon
Mathematician (algebraic geometry, motives & friends, singularities in statistics and ML). 'Geometry is successful magic' (R. Thom) University of Amsterdam.
你可能會喜歡
"Pessimism of the Algebra, optimism of the Geometry"
My coworker showed this to me and I've been cackling since
OH NO HE DIED
What’s a line delivery you’ll never forget?
Wow, someone in the Trump administration blatantly insider trading Swamp drained libtard
SOMEONE JUST OPENED A #BITCOIN SHORT 30 MINS BEFORE TRUMP'S TARIFF ANNOUNCEMENT AND JUST CLOSED WITH $88,000,000 PROFIT HE OPENED THIS ACCOUNT TODAY 👀

We're like 99% sure that's not where it came from but
What if Cain actually ate Abel and that’s where the word Cannibal came from
this bothered me so much that i bought a novelty domain name and built naive sgd + a simple mlp classifier in javascript in order to try to convey the concept

i think it’s important to know that the vast majority (99.999..%) of people in the world believe that models such as Claude are intentionally programmed by humans, and that they find the supposition that they are instead grown or evolved to be absurd science fiction.
I, too, made the mistake of *checks notes* taking OpenAI's charitable mission seriously and literally. In return, got a knock at my door in Oklahoma with a demand for every text/email/document that, in the "broadest sense permitted," relates to OpenAI's governance and investors.

One Tuesday night, as my wife and I sat down for dinner, a sheriff’s deputy knocked on the door to serve me a subpoena from OpenAI. I held back on talking about it because I didn't want to distract from SB 53, but Newsom just signed the bill so... here's what happened: 🧵

🚨 What do reasoning models actually learn during training? Our new paper shows base models already contain reasoning mechanisms, thinking models learn when to use them! By invoking those skills at the right time in the base model, we recover up to 91% of the performance gap 🧵

One Tuesday night, as my wife and I sat down for dinner, a sheriff’s deputy knocked on the door to serve me a subpoena from OpenAI. I held back on talking about it because I didn't want to distract from SB 53, but Newsom just signed the bill so... here's what happened: 🧵

"I used to feel guilty in Cambridge that I spent all day playing games, while I was supposed to be doing mathematics. Then I realized that playing games is math." John H. Conway Made with #python #numpy #matplotlib
Here's a joke a Cambridge mathematician once told me: Why do algebraic topologists struggle with Bott periodicity? Because they can't count to 8.
A friend of mine won an IMO gold, went on to obtain a PhD in algebraic topology, and now works at a frontier AI lab. This guy, however, doesn't know how to do integration by parts. He knows the principle, but treats those tricks as below him. AI models today give the same vibe.
An amazing lecture for anyone who loves math! Fields Medalist Mike Freedman (known for his proof of the 4D Poincaré conjecture) talks about Perelman's proof of the 3D Poincaré conjecture at @HarvardMath University, @HarvardCMSA . Here’s the link: youtube.com/watch?v=60X5M1…

From: Labyrinth Labyrinthos (1987), dir. Rintaro, Madhouse
It gets you funding, support, and training to do that. Alec Radford has done a lot obviously but this is kind of like a rich person laughing at someone for not having money. “So just go get a bunch of money!” Kind of silly and facile.
I said this to Alec Radford and he laughed in my face and said "what would a phd get you? just solve problems"
New research with the UK @AISecurityInst and the @turinginst: We found that just a few malicious documents can produce vulnerabilities in an LLM—regardless of the size of the model or its training data. Data-poisoning attacks might be more practical than previously believed.


“I’ll declare war on you if you don’t give me the peace prize” is an incredible bit

Are you a high-agency, early- to mid-career researcher or engineer who wants to work on AI interpretability? We're looking for several Research Fellows and Research Engineering Fellows to start this fall.

Some people genuinely believe that biological consciousness is magic and the same people will tell you with a straight face that believing in non-biological consciousness is magic
the only answer to to "why machines will never be sentient" is "because animals are magic"
“sorry for trauma dumping” bro i literally love and care for you? it’s an honour to learn about the complex and subtle ways that life has shaped you? i hope to be your trusted friend and confidant for the rest of my days?
Why do Buddhists only use shampoo? Because they recognize the pain of conditioning.
Tasmanian devils are endangered because 30 years ago one of them developed a universally fatal *contagious cancer* thats been ripping through the population ever since


United States 趨勢
- 1. Cowboys 63.3K posts
- 2. Panthers 63.2K posts
- 3. Fred Warner 2,662 posts
- 4. Ravens 60.7K posts
- 5. Browns 58.3K posts
- 6. #KeepPounding 6,228 posts
- 7. Dolphins 43.2K posts
- 8. Colts 52.8K posts
- 9. Eberflus 8,276 posts
- 10. Steelers 60.3K posts
- 11. Rico Dowdle 7,681 posts
- 12. Drake Maye 18.8K posts
- 13. Chargers 50.2K posts
- 14. James Franklin 47.5K posts
- 15. Penn State 64.7K posts
- 16. Pickens 16.3K posts
- 17. Herbert 14.2K posts
- 18. Dillon Gabriel 3,896 posts
- 19. #FTTB 2,322 posts
- 20. #HereWeGo 5,978 posts
你可能會喜歡
-
Alexandra ✨
@EulerWept -
Ben Spitz (61/100 improv meals)
@DiracDeltaFunk -
Dᵇ(Monkey) Alex Scheffelin
@chmonke -
Cordian Riener
@Cordian1723 -
Homie O'Morphism
@hayden_wlog -
julia🦄
@KamerynJW -
adittya chaudhuri(@[email protected])
@adittyachaudhu1 -
Aleksandra Sobieska
@combinatola -
Prakhar Pratyush 🇮🇳
@pursuingstacks -
𝕐𝕖𝕞𝕖𝕖𝕟
@yemeen
Something went wrong.
Something went wrong.