Sasha Cui

My student told me that per Yale policy, a PhD student cannot take more than 2 internships during PhD.(catalog.yale.edu/gsas/policies-…) I feel really sorry for them. Restricting their opportunities for learning real engineering expertise in industry is a pity, especially in this AI era.

@LocasaleLab Various examples come to my mind too.

I knew a student once - I was trying to get him to graduate with his PhD around his fifth year after he published a paper. I think he insisted on staying for years after. I remember he seemed to cycle through new undergraduate girlfriends with each graduating class.

@FangYi11101 @mean_field_zane Many many cheaters.

@mean_field_zane Duke for today

Interviewing a bunch of new grads this week and every resume has a 4.0 GPA. Complete noise at this point.

@jenzhuscott Can you provide evidence that Terry exhibits genuine creativity?

Agree w Terence Tao - LLMs limitations are structural. I’ve always said the usefulness of current AI correlates w the users expertise. So the illusion of creativity can impress/fool non experts. The current LLMs excel at Keplerian work (empirically testing many combinations via brute/compute scaling) but not Newtonian unification or genuine leaps. They act as a “super-assistant” for literature search, candidate generation, formalization, and exposition - freeing us for the creative core - but there is no evidence yet of autonomous originality at the frontier. Solving a Millennium Prize problem de novo w a genuinely novel technique (not latent in the corpus) would constitute such evidence; it has not occurred.

@tunguz Is there any evidence that Terry exhibits genuine creativity?

I strongly agree with this assessment.

@ProGof39474 @alz_zyd_ Totally right. That’s why the stats department is left behind. Serves it right.

@alz_zyd_ They renamed Type I and Type II errors to something less moronic. Statisticians at that time also basically refused to do any kind of data integration or cleaning and considered it beneath them. They had this coming.

A decade or so ago, computer scientists started invading statistics with their fancy "machine learning" stuff. This pissed off the statisticians to no end, since the CS guys renamed everything, ignored most of the statisticians' work, etc.

@AlexKontorovich Without the AI companies, those mathematicians will forever belong to the fringe

For a counterpoint, here is Johan Commelin, Director of the Mathlib Initiative (in the same Zulip thread):

@mean_field_zane These people tend to fail spectacularly in their careers because they’re not obsessed with their own work.

Met a really beautiful math PhD student at a house party last night. My friend and I tried talking to her, she immediately said “I don’t want to talk about AI or math” and fled the party. Why were you there if you didn’t want to talk about anything real, it was a party of econ graduate students lol.

Arguing that Elon Musk’s success is due to “narrative control”, luck, or riding others coattails is such an implausible claim that it functions as a useful litmus test for a persons analytical judgment. This isnt about whether you like Elon Musk. I don’t know him, and I am largely agnostic about him as a person. But I do know his record as a CEO, and studying management and business strategy has been a major part of my job for the past 20yrs. From that perspective I can tell you that Musk isn’t just a good CEO. He is one of the most effective CEOs of our generation. When I hear people write off Elon’s achievements bc someone else started these companies, it is a clear tell that they don’t understand business. Ideas are a dime a dozen. They are not what makes a great CEO. Execution is. And part of execution is recognizing a good idea when you see one and understanding how to build something around it that actually works. Tesla was months from bankruptcy when Musk took control. It’s now the company that forced every major automaker on earth to retool their entire product strategy. SpaceX was a startup that serious people in the aerospace industry dismissed as a fantasy. It now conducts more orbital launches than the rest of the world combined and has driven launch costs down by an order of magnitude. Starlink is on track to become one of the most consequential communications infrastructure projects in history. These aren’t narrative achievements. Theyre tangible businesses that work, at scale, in industries where failure is the default condition. And there’s a consistent pattern where Elon has repeatedly looked crazy, and then been right. The people who called reusable rockets a dream watched a booster fly back and land itself. The people who said a mainstream consumer EV company was impossible watched Tesla restructure the global auto industry. This is a person who has repeatedly seen something others cant see yet, absorbs the ridicule, and then builds toward it anyway. The PayPal criticism this author pushes is another perfect ex. Do you know how he became CEO? Elon identified the importance of network effects in the late 90s and realized he could take advantage of cheap capital during the internet bubble to pay users to join his network. He was labeled a lunatic. Losing money upfront to lock customers into your network is well understood now but it wasn’t back then. Confinity was forced to merge bc they couldn’t compete with it…and that’s based on Peter Thiel’s own account in Zero to One. Elon was considered reckless at the time. But he was right. And now we have people criticizing Musk’s Mars goal. But as Ben Thompson explained, Mars is the strategic North Star that forces you to radically confront the cost structure required to achieve it. Which leads you down the only path that actually scales, without settling for easier short-term solutions. If you’re serious about putting a city on Mars, full reusability is non-negotiable. And that engineering logic turns out to be what dramatically lowers launch costs. Which unlocks Starlink at scale. And Starlink creates the revenue flywheel that funds everything else. An Arianespace executive called reusability a dream in 2013 and said it was impossible. But the dream isnt the destination. It’s the constraint that forces you down the only engineering path that actually works. And it’s why SpaceX is a trillion company today. You can write off one company as luck. You can write off two as fortunate timing. But at some point the sheer weight of success across different industries and challenges stops looking like coincidence and starts looking like a big flashing signal. When someone executes repeatedly in industries where lack of execution destroys almost everyone else, the correct analytical move is to update your model. If you can’t see that Elon is a great CEO, then you’re just revealing the limits of your own analytical process.

Your deduplication pipeline was built for small models. At scale, it's broken. New preprint: "Scale Dependent Data Duplication" 1/10

@bartaround @KingVelesI @solanaphil but the show’s story was based on a man…

@KingVelesI @solanaphil That woman from Queens gambit series was winning everything. Clearly she was the greatest

Why are men so much better at chess than women? This is one of the most uncomfortable questions in all of gender discourse. Chess is a purely intellectual sport, but anyone familiar with it knows the performance gap is huge. In all of chess history, only three women have ever managed to crack the top 100, and as of me writing this, the current top 100 players are all men. A gap of this magnitude cannot simply be explained away by "cultural factors". In the video, you can see how the kid quickly shoots down the woman when she tries to flex being a "grandmaster". He points out that her title is actually that of "woman grandmaster", which is a different and much less prestigious title that was created precisely because so few women manage to reach the grandmaster title. So what is the reason for this huge disparity in a purely intellectual sport? I think it is quite obvious that, over the course of human evolution, men and women were naturally selected for different tasks, and that this has resulted in them developing different adaptations, both physical and mental. The gender that was constantly busy with hunting and warfare naturally developed a better ability for deep focus and tactical thinking, while the one that spent most of its time in childbearing and childrearing unsurprisingly became better at multitasking and empathy. Yes, this is a simplified look at a much more complex subject, but by and large, it holds true. Unfortunately, acknowledging this kind of realities is seen as "misogynistic" today, as our postmodern culture believes the ability to play puzzle games well is more impressive and worthy of admiration than high-level mothering instincts and skills. Today, the most underrated kind of excellence is the silent genius of an amazing mother.

Sufficiently advanced agentic coding is essentially machine learning: the engineer sets up the optimization goal as well as some constraints on the search space (the spec and its tests), then an optimization process (coding agents) iterates until the goal is reached. The result is a blackbox model (the generated codebase): an artifact that performs the task, that you deploy without ever inspecting its internal logic, just as we ignore individual weights in a neural network. This implies that all classic issues encountered in ML will soon become problems for agentic coding: overfitting to the spec, Clever Hans shortcuts that don't generalize outside the tests, data leakage, concept drift, etc. I would also ask: what will be the Keras of agentic coding? What will be the optimal set of high-level abstractions that allow humans to steer codebase 'training' with minimal cognitive overhead?

Did you know that Grok is MUCH MORE HONEST than GPT & Claude? Introducing PieArena: Frontier Language Agents Achieve MBA-Level Negotiation Performance and Reveal Novel Behavioral Differences. Deepdive: open.substack.com/pub/sashacui/p… Preprint: arxiv.org/abs/2602.05302 #AI #Agent #LLM

@JasonBotterill Europe doesn’t even have a strong Internet company.

How did Europe completely fail to have any stake in the AI race post-chatgpt? They had 3 years to mobilize and all we have is Mistral and thats basically a joke

@andrey_kurenkov Agreed. It’s over hyped and the main points were understood way back.

tbh don't get the hype about this paper, it really just formalizes / creates new terminology for the concept of sub agents + treating part of the context for a given task as readable files (i.e. What coding agents already do), no?

We acknowledge the guidance and generous financial support of Jasjeet Sekhon and Bridgewater Associates.

Introducing Painless Activation Steering (PAS), an automated, lightweight way to post-train LLM behavior post‑training. PAS boosts behavior metrics painlessly. Deep‑dive: open.substack.com/pub/sashacui/p… Preprint (arXiv): arxiv.org/abs/2509.22739 #AI #LLM #NLP

Joint work with Zhongren Chen

Mathematics and mathematicians get everywhere. Like, everywhere. Andrew Wiles @OxUniMaths sampled by rapper J Cole. Fermat's latest theorem.

Smarter models are no longer the barrier to deploying AI for real problems in enterprise - context is. Your AI is smart enough to be a teammate, it just doesn’t know what it means to do a good job. I’m excited to be supporting @workfabricai led by @RohanMurty and @gnychis!