Baker estimates Anthropic could do $100-200B in unconstrained revenue if compute weren't limited.
“So you might be buying it at more like five times unconstrained. I'm gonna make up a new number. URR, unconstrained revenue.”
Baker estimates Anthropic would do $100-200 billion ARR unconstrained by compute, valuing it at 5x URR.
“I think they would be doing materially more a hundred, one hundred and fifty, maybe 200,000,000,000. So you might be buying it at more like five times unconstrained.”
Baker estimates Anthropic could be doing $100-200 billion in revenue if unconstrained by compute availability.
“So I think they would be doing materially more a hundred, one hundred and fifty, maybe 200,000,000,000. So you might be buying it at more like five times unconstrained.”
Baker argues silicon will eat the world because AI is far more computationally intensive than deterministic software.
“And the nature of AI just as AI eats the world, silicon is going to eat the world because it's so much more computationally intensive than deterministic software written by humans.”
Baker explains AI recomputes answers probabilistically each time, enabling superhuman capabilities but requiring extreme computational expense.
“AI, even if you put a harness on it, even if you do the chain of thought, even if you have multiple agents, it's probabilistic that it is recomputing the answer each time.”
Baker notes compute brought on in June may show Q3 acceleration
“@armandorvila They brought on a lot of compute in the month of June so may be more of a Q3 acceleration there - we shall see very shortly”
“A series of fairly wild thoughts from Dwarkesh. I thought I was bullish but even I am not assuming that the price to rent compute continues to go up! I will say that I am in SV this week and Dwarkesh is capturing the zeitgeist. Net, net public markets would probably be trading https://t.co/PzIxxnzbiS”
“@jukan05 4 gigawatts this year and another 4 next year per this report. Spot price for compute has gone up since they signed their Anthropic and Google deals.”
Baker reports Funder AI believes SpaceX will attempt to bring on eight gigawatts of compute.
“a, you know, a substack writer will fund a fund a AI. They think that SpaceX is gonna try and bring on eight gigawatts of compute.”
Baker notes SpaceX added massive compute capacity instantly without any market price impact.
“they put a vast amount of compute into the market overnight, and it wasn't even really a blip. It was like the market just utterly absorbed it.”
Baker observes that contracted compute capacity is trading at a large discount compared to current spot market prices.
“Essentially, you have the contracted base of installed compute trading at a massive discount to the current spot market,”
Baker questions whether AFD is compatible with data locality, disputes OpenAI bet against disaggregation
“@maxkan Hmm. I might ask @dylan522p if Attention-FFN disaggregation (AFD) is compatible with data locality and not moving the KV cache. That constraint is what OpenAI focused on with Jalapeño. They absolutely did not make an “explicit bet that disagg is not the way to go.” There are https://t.co/2lVjH2M5vS”
Baker hypothesizes tomorrow's Ultrafast will be GPUs or Jalapeño with CS-4 or CS-5
“@dylan522p @maxkan And they also said this followed by “Cerebras FTW.” I would be surprised if their head of Codex was unaware of their roadmap. Reasonable to hypothesize tomorrow’s Ultrafast will be either GPUs or Jalapeño working with CS-4 or CS-5. We shall see! https://t.co/QM0vs809x7”
Baker speculates Ultrafast will use GPUs plus CS-4/5, notes Jalapeño not financeable at GPU rates
“@dylan522p @maxkan Maybe GPUs plus CS-4/5 for Ultrafast at scale. Jalapeño not financeable at comparable rates to GPUs and probably not that big of a wafer allocation next 2 years. Jalapeño still a really impressive achievement from OpenAI. I remember being sad when they hired Thomas away from https://t.co/a0jIXXahXN”