David Sacks

there's a specific kind of moment in washington when a policy fight and a market fight collapse into the same argument — and david sacks is right in the middle of one of those right now.

sacks grew up in south africa, came up through stanford, and landed in the early paypal orbit at exactly the right moment — that group of founders and operators who'd go on to fund and build half of silicon valley. he was COO there, then built yammer, sold it to microsoft for over a billion dollars, moved into venture with craft ventures, and became one of the more thoughtful voices in the all-in podcast ecosystem — a show that, whatever you think of it, has an outsized effect on how a certain class of operator thinks about technology and power. when trump came back to the white house, sacks was named AI and crypto czar, a role he held until stepping down earlier this year. so he's operating now in that particular post-government zone where your credibility is high and your constraints are suddenly much lower.

the reason he's where he is right now comes down to a fight over open-source AI that got very sharp, very fast. anthropic — the safety-focused lab behind claude — made a case to washington that open-weight AI models with chinese origins should be restricted or banned outright, on national security grounds. sacks pushed back hard and publicly. his argument: anthropic paid 1.5 billion dollars to distill its own models, meaning it trained on the outputs of competitors, and is now trying to use regulatory leverage to shut down the very competitive dynamic that made that possible. in his framing, what anthropic is calling a safety concern looks a lot like a moat protection strategy dressed in national security language.

he's also been tracking the chinese side of this more carefully than most. when kimi k3 — a model from the chinese AI lab moonshot — had a breakout moment, sacks was one of the first official voices to flag it as genuinely concerning, saying it was exactly the kind of development that should change how the US thinks about the pace of this race. at the same time, he's been arguing that the US is tying itself in knots — layering on restrictions, treating open-source development with suspicion — while china moves fast and publishes openly. it's a tension he seems genuinely frustrated by, not just strategically positioned around.

what makes this interesting is that sacks is holding two things at once that don't always sit comfortably together: a real hawkishness on chinese AI capabilities, and a real skepticism about using regulatory fear to kneecap american competitors. that's not a contradiction exactly, but it's a harder line to walk than most people in this conversation are trying to walk. most people pick a side — national security hawk or free-market open-source advocate. sacks is insisting both things are true simultaneously, which is either clarity or it's a tell about where his investments sit, depending on how charitable you're feeling.

what you're watching, ultimately, is someone who spent six months in the room where these decisions get made, now standing just outside it — with enough institutional knowledge to be taken seriously and enough freedom to say things an active official couldn't. that's a particular kind of leverage, and sacks knows how to use it.