David Sacks explains reports of China restricting overseas access to its top models fit a broader pattern where Chinese labs (ByteDance, Alibaba's Qwen, Zhipu's GLM) go open only until nearing the frontier, then shift closed — mirroring OpenAI's own earlier pivot. ✦ AI generated
David Sacks · All-In Podcast · 2026-07-11 · original ↗
starts at this moment · 56:38
“So explain the game on the field here. If you're going to look into what China's thinking, why would they want us to not have those open source models and how is this chessboard development?”
They I think are going closed too after having been open. And so this is I think the tactic is you stay open until you catch the frontier or you get close to it and then there's a really compelling incentive to go close because you want to capture all the value for yourself.
verbatim transcript · starts at 56:38
56:38understand it, is bite dances model, which is already closed. That's kind of like their chat GPT equivalent and it's always been closed. Then you've got Alibaba's Quen, which was open and now I think is going closed, and ZepO, which has GLM 5.2, too, which we've talked about a couple weeks ago because it seemed to be catching up to what was then commercially available as the American frontier at certain tasks. They
57:07I think are going closed too after having been open. And so this is I think the tactic is you stay open until you catch the frontier or you get close to it and then there's a really compelling incentive to go close because you want to capture all the value for yourself >> which by the way is exactly what Sam Alman did famously at OpenAI. Not only did they go from a nonprofit to a
57:28for-profit, they went from open models to closed models. So, it's exactly paralleling what Sam realized 3 years ago. Yeah. >> I mean, in a way, that was what I think Meta's original strategy was was that Llama was going to be open, but then they actually they've sort of backed away from open a little bit. But this is kind of an obvious strategy, right? is is that if you want to catch up, you go
57:52open. Because by the way, you're not going to make any meaningful revenue on on closed anyway because you're not close enough to the frontier. So why would anyone buy your product, but if you go open, you get the developer community on your side. So >> and you get utilization, more people use it, which in AI gives you reinforcement learning. Yeah. Sex. >> Well, having having spent some time in
58:10DC this week, I and and and talking with both the White House and Treasury, etc. on this topic. What I can tell you is while there may be some, you know, debates about regulation of US models, the one thing there's absolute agreement on is doing everything to stay ahead of China. And, you know, and and the president all the way up to the president very interested how far are we
58:33ahead of China? What are the things we need to do to stay ahead of China? It is a unifying force in Washington. And the idea that we were going to kind of take our frontier labs off the field, off the playing field while letting Chinese open source models run free, you know, and on top of that distilling our models. I will tell you GL GLM 5.2 has watermarks