Chinese models can match Western frontier models in narrow areas like coding because the blueprint for AI is on the internet; what China lacks is compute, so it specializes — which means model intelligence is commoditizing and the value in AI is shifting to whoever builds the best products, not the best models.
Alex Cantz explains that distillation of American models is over-credited — the real recipe is compute + data + a big model, and China's compute constraint forces specialization (DeepSeek on reasoning, Kimi on coding). Since several labs now reach near-parity, model margins collapse (OpenAI cutting prices 20-80%) and value shifts to AI products, not the models themselves. ✦ AI generated
Alex Cantz · The Compound · 2026-07-31 · original ↗
starts at this moment · 22:31
“how much of this stuff that we thought would be profitable it just turns out can be imitated really cheaply and people won't really care what they use as long as it doesn't cost too much.”
Okay. So, first of all, on this idea that the Chinese models are distilled from American models to a degree, yes. What they'll do is they'll just run a bunch of queries, get the answers, and they'll sort of bake that into the intelligence of the model. But the distillation part of this has been given too much credit because at the end of the day, what's happening with AI is the blueprint to build these models are on the internet, right? So basically all you need is compute, data and a big model and the bigger all three of those get the better performance you have. Now there are some new tricks that you can use to make the model perform better but that's at the core of this. So uh you know this idea that China couldn't build the model uh you know of course they can the blueprints are on the internet. Now the constraint within China is compute because we will not sell the cutting edge Nvidia chips to China. So what China has to do is specialize. So, you know, large language models, they're large. Um, which means that when you, you know, write a query to chat GPT, you're getting something that probably is, you know, can answer your your health questions at a high level, can answer science questions at a high level, uh, can go search the internet and tell you if your train is delayed, uh, can go into your email and draft emails for you. Uh, what's happened in China is there's been a specialization, right? a constraint and this is something Grace Shiao who's a a China analyst uh told me on the show this week on my show this week is basically because China is constrained they have to focus all their efforts on certain areas so deepseek was all about reasoning which is one of these practices to make the model better and uh and Kimmy K2 has been about this agenda coding so it might not do so well on health like an open AI model does but in areas like coding they can equal the frontier so that's why we're seeing this challenge
verbatim transcript · starts at 22:31
22:14really care what they use as long as it doesn't cost too much. >> Okay, so there's so much here. Let's go bit by bit. All right. >> You know, you know this really well, so we need to learn this. >> Okay. So, first of all, on this idea that the Chinese models are distilled from American models to a degree, yes. What they'll do is they'll just run a
22:31bunch of queries, get the answers, and they'll sort of bake that into the intelligence of the model. But the distillation part of this has been given too much credit because at the end of the day, what's happening with AI is the blueprint to build these models are on the internet, right? So basically all you need is compute, data and a big model and the bigger all three of those
22:51get the better performance you have. Now there are some new tricks that you can use to make the model perform better but that's at the core of this. So uh you know this idea that China couldn't build the model uh you know of course they can the blueprints are on the internet. Now the constraint within China is compute because we will not sell the cutting edge Nvidia chips to China. So what
23:13China has to do is specialize. So, you know, large language models, they're large. Um, which means that when you, you know, write a query to chat GPT, you're getting something that probably is, you know, can answer your your health questions at a high level, can answer science questions at a high level, uh, can go search the internet and tell you if your train is delayed, uh, can go into your email and draft
23:35emails for you. Uh, what's happened in China is there's been a specialization, right? a constraint and this is something Grace Shiao who's a a China analyst uh told me on the show this week on my show this week is basically because China is constrained they have to focus all their efforts on certain areas so deepseek was all about reasoning which is one of these practices to make the model better and
23:57uh and Kimmy K2 has been about this agenda coding so it might not do so well on health like an open AI model does but in areas like coding they can equal the frontier so that's why we're seeing this challenge >> because they just have this one narrow thing that they're perfect for. >> Put your smartest people on this one thing. >> So there So there is a a website is it a
24:16what is it? AI router or >> you know >> somebody's trying to buy it for $10 billion right now. >> Right. Right. Right. >> This is where you go to it and you say what you want done and it's deciding which model or which combination of models >> can you throw a harness around and have them pull together and come up with the right answer. It sounds like if there
24:37are going to be a lot of specialized models and you're a specialist in a field, right? If you're >> Oh, so that's open router. >> Yeah. So, if you're like if you're doing drug development, you're probably not on chat GPT. You probably already have a favorite specialty. Uh but like having this sort of like the Google of routing AI queries is an interesting thing. So much so somebody's trying to buy a
- ·AI blueprints are on the internet — compute, data, scale win
- ·Distillation over-credited; the real recipe is well-known
- ·Multiple labs now reach near-parity with frontier models
- ·China lacks cutting-edge Nvidia chips
- ·Must focus: DeepSeek on reasoning, Kimi on coding
- ·Can equal frontier in narrow domains despite constraints
- ·Near-parity collapses model margins
- ·OpenAI cutting prices 20-80%
- ·Advantage moves to best products, not best models