What Dylan Patel thinks about LLMs
Founder and chief executive of SemiAnalysis, the research firm that tracks the semiconductor and AI data-centre supply chain.
Dylan Patel did not write this page.
We collected these quotes from things they published elsewhere, and every quote links to where it was said. They have no account here and have not endorsed this site. Quotes are word for word; the short line under each one is our own restatement, not their wording. Their own site. Is this you? Claim it or ask us to remove it. Or tell us what is wrong here.
Filter LLMs · Newest first
29 more topics
Only Dylan Patel on LLMs. See everyone on LLMs
15 dated positions, 2025 to 2026, in their own words. Our reading of what Dylan Patel has said — not written or endorsed by them.
-
Our readingAI labs' cash flows cannot yet fund the well over $2 trillion of capex the build-out needs.
Their wordsYou sum all this up, it's going to be well north of $2 trillion of CapEx. So the labs have not yet gotten to the point where their cash flows can fund this stuff.
↗Dylan Patel – Two labs will soon control most of the world's workforceyoutube.com 7th of 30 in this recording
-
Our readingThe regulation the AI labs advocate slows them far more than it slows Chinese open-source models.
Their wordsThis regulation that they advocate for is actually slowing down the labs a lot more than it slows down, you know, sort of the open -source Chinese language models.
↗Dylan Patel – Two labs will soon control most of the world's workforceyoutube.com 8th of 30 in this recording
-
Their wordsThey're clearly not releasing their best models. And in which case, their revenue per megawatt stalls or even can start to decline again because other models are competitive again.
↗Dylan Patel – Two labs will soon control most of the world's workforceyoutube.com 9th of 30 in this recording
-
Their wordsLike most of the value that these models generate does not get given to OpenAI Anthropic. Thankfully, so far, it is mostly just being given to the users, right?
↗Dylan Patel – Two labs will soon control most of the world's workforceyoutube.com 10th of 30 in this recording
-
Their wordsI personally believe that the labs are going to allocate less and less compute to inference over time, which I think is very non -consensus, right?
↗Dylan Patel – Two labs will soon control most of the world's workforceyoutube.com 13th of 30 in this recording
-
Our readingThe only reason for labs to keep inference compute so large is to grow their training fleets.
Their wordsultimately it makes more sense to dedicate more and more compute internally. And the only reason to have inference compute be so large is so you can grow your training fleet.
↗Dylan Patel – Two labs will soon control most of the world's workforceyoutube.com 15th of 30 in this recording
-
Their wordsChinese companies today are not that far behind in AI models, at least perceivably by the public, relative to the amount of compute they have, right?
↗Dylan Patel – Two labs will soon control most of the world's workforceyoutube.com 21st of 30 in this recording
-
Their wordsAnd that is actually, like, 50 % of the compute is research, like, 10 % of the compute is development, and then 40 % is inference.
↗Dylan Patel – Two labs will soon control most of the world's workforceyoutube.com 22nd of 30 in this recording
-
Their wordsBecause it seems to me that every force is screeching toward centralization. Right. And that's scary as hell.
↗Dylan Patel – Two labs will soon control most of the world's workforceyoutube.com 28th of 30 in this recording
-
Their wordseverything along the way will still lead to someone's going to allocate, going to capture more resources. So, it's kind of hard for a framework in which AI doesn't lead to super concentration.
↗Dylan Patel – Two labs will soon control most of the world's workforceyoutube.com 30th of 30 in this recording
- 19 months earlier
-
Their wordsI think it's even more impressive what OpenAI did in 2022. At the time, no one believed in mixture of experts models at Google who had all the researchers. OpenAI had such little compute and they devoted all of their compute for many months, all of it, 100% for many months to GPT-4 with a brand-new architecture with no belief that, "Hey, let me spend a couple of hundred million dollars, which is all of the money I have on this model." That is truly YOLO.
↗DeepSeek, China, OpenAI, NVIDIA, xAI, TSMC, Stargate, and AI Megaclusters | Lex Fridman Podcast #459youtube.com 4th of 44 in this recording
-
Their wordsTo some extent, training a model does effectively nothing. They have a model. The thing that Dario is sort of speaking to is the implementation of that model, once trained to then create huge economic growth, huge increases in military capabilities, huge increases in productivity of people, betterment of lives.
↗DeepSeek, China, OpenAI, NVIDIA, xAI, TSMC, Stargate, and AI Megaclusters | Lex Fridman Podcast #459youtube.com 7th of 44 in this recording
-
Their wordsOpenAI has a fantastic margin. When they're doing inference, their gross margins are north of 75%. So that's a four to five X factor right there of the cost difference, is that OpenAI is just making crazy amounts of money because they're the only one with the capability.
↗DeepSeek, China, OpenAI, NVIDIA, xAI, TSMC, Stargate, and AI Megaclusters | Lex Fridman Podcast #459youtube.com 17th of 44 in this recording
-
Their wordsBut the funniest thing I think that comes out of this is Jevons paradox is true. AWS pricing for H100s has gone up over the last couple of weeks, since a little bit after Christmas, since V3 was launched, AWS H100 pricing has gone up.
↗DeepSeek, China, OpenAI, NVIDIA, xAI, TSMC, Stargate, and AI Megaclusters | Lex Fridman Podcast #459youtube.com 26th of 44 in this recording
-
energysemiconductorsnuclear powerJapan
Their wordsSo, Japan has a law which you're allowed to train on any training data and copyrights don't apply if you want to train a model, A. B, Japan has 9 gigawatts of curtailed nuclear power. C, Japan is allowed under the AI diffusion rule to import as many GPUs as they'd like.
↗DeepSeek, China, OpenAI, NVIDIA, xAI, TSMC, Stargate, and AI Megaclusters | Lex Fridman Podcast #459youtube.com 31st of 44 in this recording