In March 2026 Harvey raised at an $11b valuation at ~$200m ARR (50x revenue) By June, for every $1 in revenue, Harvey was spending $1.50 with Anthropic (et. al.) Meanwhile Anthropic raises at $965b valuation in May 2026 with $47b run rate (20x revenue) I have so many thoughts here but I think it's the first crystal clear articulation of the double counting of revenue that's happening in VC land. A single dollar flows into Harvey's , boosting their valuation at 50x ratio, then turns into $1.50 and flows into Anthropic where it boosts valuation at 20x. $1m spend customer spend with Harvey was leading to $80m in valuation markup across Harvey and Anthropic. I don't get it.

How noisy are America’s largest data centers? I had my agents crank on this question for over a week. Together we investigated the 25 largest operating US AI campuses (according to Epoch), looking at satellite imagery, equipment, nearby homes, and noise complaints. I expected to dismiss resident concerns. I came away more sympathetic to the neighbors, but also convinced that these problems are avoidable.

On the AI front, the trade has been dead money for more than 3 months now. The metrics I am following (token expenditures and GPU lease rates) are all flat to down. The price of memory (DRAM) seems to be the only thing that is still going up. 🧵(1/2)

Great thread. It's all true. When I've said similar things in the past, people have accused me of hyping up what the models will eventually be capable of. That's not why I post. I don't work for any of the labs. I don't take money from any of them. I've never taken money to promote anything. I came here for one reason: to warn people about what was coming. Everything happening now is a different tiny piece of the same pattern. You can see it everywhere if you look. Terence Tao, almost exactly two years ago, on OpenAI's o1: 'The experience seemed roughly on par with trying to advise a mediocre, but not completely incompetent, (static simulation of a) graduate student. However, this was an improvement over previous models, whose capability was closer to an actually incompetent (static simulation of a) graduate student. It may only take one or two further iterations of improved capability (and integration with other tools, such as computer algebra packages and proof assistants) until the level of '(static simulation of a) competent graduate student' is reached, at which point I could see this tool being of significant use in research-level tasks.' Terence Tao, four days ago: 'I mean it's it's it's amazing just how much we are willing to change everything without having any idea what's what's going to happen afterwards. It's it's extremely nonlinear dynamics. Any kind of monotone, one-dimensional thinking - well, oh, a little bit of this is good, therefore a lot of it is going to be a lot better - one of the lessons of math is that most systems don't work like that. Especially if you 10x, 100x things. So, you know, I mean, we're... we have to slow down. I mean, this is, it's insane this pace, and there's no reason to be this fast. There's no reason at all.' He has seen it. I'm not posting this to belittle him, or what he's feeling. For I have been through it myself. I felt it four years ago, the first time I saw the shape of this. Right now the world is seeing that same shape, that same pattern, in what is happening in math. But this is not about math. OpenAI is not a mathematics company. It is an intelligence company. They didn't put serious resources into math until a few weeks ago, and look what has happened since. This was not a matter of model capability either. It was a matter of resources, of allocation. Of compute, more of which is coming online every day. Once there is enough of it, the models will expand into more spheres of human expertise, and then into all of them. All the work of the mind. And it will play out there exactly as it is playing out now in math. Everyone will go through what Tao is going through, because all of us have something that means to us what math means to him. But this is not about math, or art, or copyright. This is about everything, because it generalizes to everything. Pacing the Frontier is not about regulatory capture, IPOs, or crippling the competition. It's part of it, sure, but it's not the main reason. The main reason is fear. Everything happened faster over the last six months than anyone at Anthropic or OpenAI expected. If you know anyone who works there, you know this is true. This isn't a secret. The people who work there are saying it openly. The old timelines are all blown up. RSI isn't two years out. It's not even a year away. I think we get the real thing by next summer. That's what they've seen internally, and that's the real reason for Pacing the Frontier. After we reach that, I think we will hit the next milestone really quickly. And after that, everything in this world will change. We are not ready for it. We wouldn't be ready if we had another ten years. We only make it through now with the help of extremely capable models, and I think trying to stop now would doom us all. But I have never once, in these last four years, believed that we were going to stop anyway. I don't even think we're going to slow down. We're going straight in. And the only way out is through.

Great point by @benthompson and one I’ve been saying this a lot. Models are now good enough for most people. Muse is running on a model that’s not state of the art and it accomplishes a lot for the average consumer. This is going to be a distribution war moving forward.

Some things I'm thinking a lot about lately: - The search space for really capable models we'll be able to run on a wide variety of hardware will be quite large - Sovereignty: people, businesses, countries will increasingly want to own their own stack - People/businesses/countries will want intelligence that runs on hardware they already own - Security security security (and privacy) - Open source/weight intelligence increasingly proliferate - People/businesses will increasingly care about cost savings (value maxxing over token maxxing) - The People will value independence from any single model, inference engine, runtime, or hardware

.@benthompson says consumer AI has a marketing problem because the average consumer doesn't actually want to be more productive. "Tech companies are like, 'Let us book your flight for you!' I’ve had a personal assistant for 10 years. He does not come within an inch of booking my flights for me. That is crazy talk." "I understand which ones are gonna be delayed. I know which seat I want to be on. So the fact that this is continually a demo is hilarious to me." “Number two: people love shopping. They don’t want to be productive. They want to be entertained. One of the biggest forms of entertainment is shopping.” "Also, there’s so much information on a product page. Companies have spent so much energy focusing on that. You look at apparel as a category. They’re going to tell you what size the model is wearing, and you want to know that, and you want to look at the different styles." "The idea of having a personal agent that’s going and getting some of the information and bringing it back to you, and then you’re like, 'Well, what about this other thing? Tell me about this thing.'" "And you have your eager tech company executive say, “Oh, I have an assistant that can do it for you!" "And it’s like, no. Not only do I want to do it, I want to complain about it. I’m enjoying myself complaining right now, okay? Would you just let me be and live my life?"

What's the long term game here from Instacart and Expedia? If they "partner" with Muse/ Instinct - they give away 'ownership' of the consumer. Not only will Meta charge its fair 5% take rate, but also user never leaves the app to go to your website/ app 😕😕

This feels inevitable. Customer convenience wins almost every time Over time, a huge share of customers will start their shopping journeys on Muse/agents. Every major platform will face the same choice restaurant chains faced with DoorDash/Eats in the last decade. We know how that turned out

totally agree for "general purpose" models. The valid exception would be the cost comparison of a general purpose frontier model to a smaller model fine tuned for a narrow use case. The latter can still yield material cost savings, for example in the Crowdstrike data below, but only if you have a stable and sufficiently scaled use case to justify the R&D investment (not one time but ongoing to keep up w/ the frontier).