Dilip Krishnan

Earlier this month I volunteered at Stanford’s Future of Math symposium, and ever since, I've been puzzling through what it now means to pursue mathematics as a student in the age of AI. I wrote an essay to make sense of it all: apoorvapanidapu.substack.com/p/a-new-consci…

A Buddhist hall in Western Japan known for housing an "eternal flame" that has been burning for over 1,000 years caught fire and burned to the ground on Miyajima Island.

Whoa. This breakthrough is going to fundamentally affect the structure of how universities select and retain professors. And more generally the structure of work and creativity. AI has managed to discover a significant result in research mathematics in a one-shot query (!!!), disproving a conjecture that a huge number of people have tried working on (including myself, although I am not anywhere as good as the other experts who have thought about it). @wtgowers (Fields Medalist who has been thinking a lot about this space of AI and math) wrote: "if a human had written the paper and submitted it to the Annals of Mathematics and I had been asked for a quick opinion, I would have recommended acceptance without any hesitation. No previous AI-generated proof has come close to that." [The Annals of Mathematics is perhaps the most prestigious math journal in the world.] OpenAI's announcement: openai.com/index/model-di… Mathematicians' discussion: cdn.openai.com/pdf/74c24085-1… In college classes, we already have a major issue where take-home assignments are susceptible to students using AI. Even with in-person exams, we're having issues where students use AI in the bathroom during the exam. Now, how will universities decide which professors to hire and promote? Universities used to judge on the basis of whether you got papers published in highly prestigious journals. Should the person who can use AI to generate massive quantities of publishable results be picked, if it can be done with one-shot prompts? More generally, everyone (not just universities) needs to rethink their objectives for hiring and promotion. I am also an entrepreneur, and have been refining my hiring process in this age of AI. Before, I used to be particularly impressed by academic competition performance. Today, I search for people who hold 2 certain principles very strongly: they enjoy (1) delighting other people, and (2) achieving understanding through their own thought. I find these people are good at figuring out what makes customers/partners tick (hence able to identify good directions to run without micromanagement), and also curious enough to learn forever. They also tend to already be pretty strong skills-wise, because those two principles inherently drive them to build skills. I actually think the whole world would be better off with more Thought Full people: thoughtfull.net. We'll need people like that to figure out how to help humanity survive. If you'd like to collaborate on ways to make a future, feel free to reach out. That's all I work on nowadays.

this is what children in shenzhen learn about in their science and tech museum: - supply chain logistics - photolithography for chip design - applications of mxene-liquid crystal elastomer materials (in solar/optics/robotics) - biological 3D printing just to name a few... so what will you (and/or your children) learn about today?

It was very fun to hear Ron tell the stories of early Jane Street. Their first cluster was a pile of 6 Dell boxes in their office. And it was important to them that it be physically in their office so that if something went haywire, they could just physically unplug the machine. Goes without saying that they can no longer fit their 100s of k of GPUs in their office.

Can we all agree that LLM-powered hyper param search to optimize nanoGPT better is not really AI research?

Krishna Rao is the CFO of Anthropic, and this is his first podcast appearance. He joined the company two years ago when run-rate revenue was about $250M. Today it is $30B. He has helped raise ~$75B and is responsible for the procurement and allocation of compute. I feel lucky we get to hear what it is like to sit inside a company this consequential at a moment this pivotal. We discuss: - The cone of uncertainty - How he allocates compute across Trainium, TPUs, and GPUs - What investors misunderstand about model companies - Why the returns to frontier intelligence keep rising - Platform vs application and where Anthropic builds its own products - How Anthropic uses Claude internally I have asked my closing question about the kindest thing more than 500 times. Krishna's answer is one I have never heard before. Enjoy! Timestamps: 0:00 Intro 2:38 The Compute Canvas 6:51 The "Cone of Uncertainty" 11:58 Why the Returns to Frontier Intelligence Are So High 16:45 Recursive Self-Improvement 20:20 Scaling Laws 23:30 Sourcing $100 Billion in Compute 28:05 Platform vs. Application Strategy 32:52 Pricing Dynamics 38:48 How Anthropic’s Finance Team Uses Claude 43:24 Raising Capital & Overcoming Investor Skepticism 52:32 Public Perception, Risks, and Government Regulation 57:25 Mythos Release 1:12:33 What Could Derail the AI Revolution? 1:13:47 Biotech and Healthcare 1:15:31 The Kindest Thing

Excited to co-found Recursive (@recursive_si) with an exceptional team in London and SF to create AI that experiments on how to safely improve itself, turning compute into knowledge that accumulates in an open-ended process of endless, automated scientific discoveries.

Unitree Unveils: GD01, A Manned Transformable Mecha, from $650,000 👏 The world's first production-ready manned mecha. It can transform. It's a civilian vehicle. It weighs ~500kg with you inside. Please everyone be sure to use the robot in a Friendly and Safe manner.

@_arohan_ A bit, and yes.

What research did you get done this week? Was it directionally correct?

If the AI models are so smart, why do I feel like I’m losing a few neurons every time I read a longer form content written by AI? We’ve come a long way but we still have long way to go. In terms of clarity of writing we may have regressed from o1/o3 days.

The Indian Institutes of Technology (IITs) in india have the most objective selection process of any undergraduate engineering program worldwide. There are no essays or legacy priorities. One just has to do exceedingly well on two ultra tough tests. The IITs admit just 1.1% of applicants! It is 3-4x easier to get admitted to @mit, @Stanford, @Harvard and @Princeton! I am excited to share that this year one of our @dakshanaindia Scholars, Godavarthi Harshit Visweswar, got a rank of 39 (out of over 1.5 million applicants!) in the first IIT selection test. He is a sure shot to be admitted and get the campus and major he prefers. His friends call him GHV and I’ll do the same. GHV grew up in the remote hinterlands of Prakasam District in the State of Andhra Pradesh in India. His father is a farmer and the family somehow survives on $5/day. When GHV was in 8th grade, a @dakshanaindia alum, Keshava Chandra, who used to be in the same school as GHV delivered Dakshana’s Inspire session at the school. Each year our alums visit hundreds of schools to inspire the next generation of students to aspire to be @dakshanaindia Scholars. GHV was impressed that Keshava was a student at @iitkgp and understood that if he wanted to go to IIT, he needed to be accepted as a @dakshanaindia Scholar after 10th grade. If there was no @dakshanaindia, the closest location for GHV to get coached for the IIT entrance exam is over 150 miles from his home and the costs of $4000+ would have been completely out of reach. Sometimes all we need is a gentle nudge in the right direction. I received a similar nudge in 8th grade and it totally changed my life trajectory. GHV was the only kid from his school accepted by @dakshanaindia in 2024. We relocated him to Dakshana’s Center of Excellence in Bengaluru and he was coached for two years at The Charles T. Munger Hall. Charlie would be proud! Congratulations to GHV, his family and his teachers. The future for GHV and his descendants looks very bright. This is @dakshanaindia at its very best!

😶mad-high investor FOMO in AI startup land doesn't hurt that @MillionInt has star studded team

Back to back rounds are continuing in neolab land! Latest example: Jerry Tworek’s Core Automation raising at a $4b valuation weeks after raising at $1b w/ @julia_hornstein theinformation.com/articles/ex-op…

Texas just passed California in utility-scale solar. And it's not close in wind or energy storage.

Taiwan economy grows 13.69% in Q1, highest in 39 years taiwannews.com.tw/news/6351770

Someone buried a 150lb chest with $10,001 of cash, somewhere within 7 miles of SF's city hall. This is the only clue. Go find it!

I came back to the United States, hearing about a very sad news: Craig Venter passed away today. He is a pioneer, successfully sequenced the first Human genome, and tried to create Synthetic Cells. We lost a giant in Science. RIP. @JCVenterInst

Did a very different format with @reinerpope – a blackboard lecture where he walks through how frontier LLMs are trained and served. It's shocking how much you can deduce about what the labs are doing from a handful of equations, public API prices, and some chalk. It’s a bit technical, but I encourage you to hang in there - it’s really worth it. There are less than a handful of people who understand the full stack of AI, from chip design to model architecture, as well as Reiner. It was a real delight to learn from him. Recommend watching this one on YouTube so you can see the chalkboard. 0:00:00 – How batch size affects token cost and speed 0:31:59 – How MoE models are laid out across GPU racks 0:47:02 – How pipeline parallelism spreads model layers across racks 1:03:27 – Why Ilya said, “As we now know, pipelining is not wise.” 1:18:49 – Because of RL, models may be 100x over-trained beyond Chinchilla-optimal 1:32:52 – Deducing long context memory costs from API pricing 2:03:52 – Convergent evolution between neural nets and cryptography

Food AI is about to have its ChatGPT moment. Our first paper is now on arXiv: Epicure. For decades, food has been treated as too human, too sensory, too cultural, and too computationally intensive to model properly. We broke that assumption. 🧵