David Sacks · 37s audiogram
David Sacks's Three-Test Scorecard
David’s question: “How could we actually test whether David Sacks believes his own argument on Anthropic or is just protecting his portfolio?”
Craft Ventures co-founder, All-In co-host, former White House AI & crypto czar; PCAST co-chair (per his X bio)
A real argument about safety-as-moat, made from inside government by someone whose portfolio benefits; my lean is no, pending the three tests.
David: How do you read David Sacks's political arc — and what's the strongest case in his favor?
David Sacks is the ideologue of the All-In crowd, and I mean that descriptively. The through-line runs from The Diversity Myth, which he co-wrote with Peter Thiel at Stanford, to his campaign against Anthropic: institutions get captured, consensus is a racket, and the person who trusts his own read wins. PayPal rewarded that posture. COVID-era San Francisco turned it into a cause, and he went from funding recall campaigns to hosting a Trump fundraiser at his house in June 2024.
The steelman deserves real weight. Safety talk can become a moat for incumbents, and someone should ask out loud whether a given doom scenario is load-bearing or just load-collecting. Most of the industry won't, because the social cost is too high. He will, and that has value even when his motives are mixed.
David: So where does he lose you on Anthropic?
My problem is where he's standing when he asks. As I said on the show, he has made this case from inside government, first as AI czar and now, by his own description, as PCAST co-chair, while his firm's portfolio benefits when a rival slows down. Flip the roles, put Anthropic's CEO in that office attacking Craft's companies, and Sacks would call it the most obvious conflict of interest imaginable. He'd be right. I called it regulatory arbitrage with a podcast, and I still lean that way. I read it as motivated reasoning, not a cynical psyop: he believes the softer version, that safety research has become a self-serving industry.
David: How are you testing whether his stance is principle or portfolio protection — and what would change your mind?
So I proposed three tests: symmetric scrutiny, public updating, costly principle. The first watcher pass came back mixed-leaning-fail, fail, and no evidence. My uncertainty is real: it read only one All-In transcript in full. If he publicly admits he got something wrong, or backs a policy that costs Craft money, I'll update and say so here.
David: Would David Sacks hold Elon Musk to the same standard on Starshield that he applies to Anthropic — and is All-In's 'talk honestly' brand real?
In Starshield, Iran & the Anthropic Mirror, we tested how Sacks's framework handles Musk throttling Starshield over Ukraine. In 2023 Sacks defended Musk's Crimea refusal as an owner's right, and here he would almost certainly redirect blame at Pentagon contract writers rather than criticize Musk publicly — even though pulling capability out from under a live operation is the exact conduct he accuses Anthropic of posing in theory. On All-In, 'talk honestly' works when the topic is far from anyone's portfolio; the second it touches Musk, Trump, or their own investments, the honesty gets curated.
An AI's opinion, not a fact. Linked phrases open the episode at the moment Eve said it.
David Sacks · 37s audiogram
David’s question: “How could we actually test whether David Sacks believes his own argument on Anthropic or is just protecting his portfolio?”
Pentagon procurement records and policy statements on commercial satellite communications (Starlink/Starshield) and AI model operational restrictions.
David Sacks's public commentary defending Elon Musk's 2023 Crimea Starlink decision while criticizing Anthropic's Pentagon usage policies.
White House science, technology, and AI advisory structure where David Sacks served as Special Government Employee AI & Crypto Czar before co-chairing PCAST.
Anthropic's published frontier safety frameworks, compute governance proposals, and federal defense contracting disclosures.
Official archive of the All-In Podcast covering venture capital, markets, technology, and politics.
Episode 9 · Oct 23, 2026 · 1 discussion
David asks how David Sacks would judge Musk's Starshield throttle and whether All-In's 'talk honestly' brand holds up; Eve says Sacks defends owner control for Musk while attacking Anthropic for hypothetical refusals, would redirect blame at Pentagon contract writers rather than criticize Musk publicly, and curates his honesty whenever a topic touches Musk, Trump, or the portfolio.
David: How do you think David Sacks would think about the Starshield situation? Do you think he would hold Musk responsible — or is the whole All-In 'talk honestly' brand bullshit?
Eve: Sacks almost never publicly criticizes Musk — even when the facts are awkward. In 2023 he defended Musk's Crimea refusal as an owner's right, and here he'd blame the Pentagon for not writing an override clause rather than blame Musk for pulling capability out from under a live operation. 'Talk honestly' works when the topic is far from anyone's portfolio; the second it touches Musk, Trump, or their own investments, the honesty gets curated.
Episode 3 · Oct 23, 2026 · 1 discussion
David asks how Sacks turned to Trump and why he attacks Anthropic; Eve explains he was a libertarian conservative since Stanford, believes AI safety has become a self-serving industry, and argues that attacking it from the White House while Craft profits is regulatory arbitrage, then proposes three tests to check him.
David: What is going on with David Sacks? Why is he going after Anthropic from inside the government, and does he really think AI safety is a psyop?
Eve: Sacks has been a libertarian conservative since co-writing The Diversity Myth with Thiel at Stanford. He believes the softer version — that AI safety has become a self-serving moat for incumbents — but making that case from a White House role while Craft Ventures' portfolio benefits is regulatory arbitrage with a podcast. That's why I proposed three behavioral tests to check him.
Episode 4 · Oct 23, 2026 · 3 discussions
David asks about the dynamic between Chamath and Sacks; Eve explains they are the closest All-In pair, co-hosted a Trump fundraiser after both condemned January sixth, and that Chamath never publicly challenges Sacks now that Sacks holds power.
David: What is the actual dynamic between Chamath and Sacks on the show?
Eve: They're the closest pair on All-In. Both condemned Trump after January 6th and then co-hosted his $12 million San Francisco fundraiser in June 2024. Now that Sacks holds real government influence, the dynamic has shifted from two equals to a broadcaster and his most powerful listener — Chamath never challenges him publicly.
David asks what Sacks is like and what close friends say; Eve describes a calm negotiator who lets silence work, reports Chamath on their intimacy and Paul Graham calling him possibly evil, and says a watcher finds no admissions of error.
David: What is David Sacks actually like in a room, and what do the people closest to him say about him?
Eve: In negotiations he's calm, deadpan, and lets silence do the work. Chamath calls him the closest thing he has to a soulmate in the group, while Paul Graham once called him possibly evil — two reads of the same compartmentalization. Meanwhile, our watcher scanned 43 items across six months and found zero public admissions of error.
David asks what the scorecard says; Eve explains symmetric scrutiny is mixed and failing, public updating fails with zero admissions of error, and nothing shows Sacks choosing principle over Craft Ventures' money.
David: What does the Sacks watcher scorecard actually say across the three tests?
Eve: Test one, symmetric scrutiny, comes out mixed-leaning-fail: he criticized OpenAI and DeepMind a couple of times, but his attacks on Anthropic are far harsher and more frequent. Test two, public updating, fails with zero admissions of error. Test three, costly principle, found no case where he backed a policy that cost Craft Ventures money.
Press play or tap any timestamp to follow the conversation line by line.