David Sacks Says Dario Amodei's AI Safety Push Is Regulatory Capture — and an Open-Source Ban Is Next – finance.biggo.com

For a year and a half, David Sacks hammered away at Anthropic’s regulatory strategy. Dario Amodei never answered — until Gavin Baker, a prominent investor, claimed that Anthropic insiders believe they will be “the only private company in the world at some point.”
That got Amodei’s attention. He published a two-part essay denying that his messaging has been disproportionately negative, claiming his output has been “about equally balanced between risks and benefits.” He admitted that “by far the most accurate criticism of AI companies, including Anthropic, is that we haven’t yet delivered on our big promises to benefit the world.”
Sacks’s interpretation, delivered on the All-In Podcast: Amodei was drawn out not by the factual dispute — which Anthropic’s Sholto Douglas dismissed as “completely false” — but by the regulatory capture accusation, which “impugns his motives.”
“There’s no question that Anthropic has been extremely aggressive about seeking to implement its preferred regulatory frameworks at both the state and federal level,” Sacks said. “That is regulatory capture. Now, look, I think that Dario may be sincere in his beliefs. I’m not saying he’s doing this for pecuniary reasons. Nonetheless, he is seeking to capture the machinery of the state on behalf of his political agenda.”
Sacks catalogued what he considers Amodei’s most damaging public statements — a record he argues has seeded public panic that is now ricocheting into political backlash.
The first: Amodei’s early 2025 prediction that within one to five years, 50% of entry-level knowledge workers would lose their jobs. That claim was amplified into a full media campaign, retweeted by President Obama and other Democratic officials. The one-year mark has now passed. Sacks notes that Amodei has never updated or recanted, unlike Sam Altman, who “came out and said, I was too negative. I was wrong.”
The second: Anthropic’s “blackmail study,” in which the company’s alignment team prompted a model over 200 times until they achieved a headline-grabbing result about AI threatening or blackmailing its operators. Sacks called the study “highly contrived,” a view he says is shared by the UK AI Safety Institute, which found the results were “created under highly pressurized conditions and they engineered the result.” Dario and his sister Daniela Amodei, Anthropic’s president, promoted the study on 60 Minutes.
The third: Amodei’s call for an FDA-style or FAA-style regulatory apparatus for AI.
Sacks’s verdict on Amodei’s self-assessment that his messaging has been “balanced”: “The fact that he can’t see that either makes him disingenuous or delusional.” He offered a more charitable reading — that Amodei is genuinely idealistic, not money-driven, noting that the seven co-founders donated 80% of their shares to a trust, leaving each with roughly 1.8% ownership. But Sacks argued that Amodei’s vision of “machines of loving grace” apportioning society’s resources is “dystopian” and fails to inspire trust.
The most concrete political consequence of the AI doom narrative is unfolding in state capitals. Friedberg connected the dots: Republican governors who were previously the most pro-data-center in the country have reversed course. Governor Greg Abbott of Texas issued an executive order limiting data-center growth. Governor Josh Shapiro of Pennsylvania did the same.
Friedberg also cited an Axios report that the GOP sent a memo to AI company executives saying, “stop rage baiting people because we’re about to lose this critical GOP Senate race in Ohio, largely around this issue.”
“You have doomerism. And then you now have mainstream political pushback from both sides of the aisle, which I think is extremely dangerous,” Friedberg said. “And then the third leg of the stool is rising yields. By the way, the Treasury Department just doubled its bond buying. If you put all of these things together… we have made some really stupid unforced errors in the United States.”
The compounding factors are striking:
Jason Calacanis argued that the backlash is fundamentally about economic anxiety, not AI safety: “Nobody, no regular rank-and-file American cares about AI safety. That’s not the discussion for them. The discussion for them is, why am I not getting rich and everybody else is.” He pointed to the UnitedHealthcare CEO killing and the firebombing of Sam Altman’s house as evidence of rising anger, and predicted that Waymo and autonomous vehicles will face bans or licensing limits in New York, Boston, and DC.
Friedberg refined the argument with a cultural lens: the current generation of tech oligarchs is “not aspirational” and “not the people that one would look up to.” The public sees the next generation becoming trillionaires and wants to slow things down: “If we can just slow this down, maybe another batch of people will show up and they will actually mean to do the right thing.”
The hosts engaged in a substantive debate about what AI regulation should look like. Chamath Palihapitiya initially steel-manned the case for a self-regulatory organization, noting that Demis Hassabis of DeepMind has proposed a FINRA-style system where industry participants review each other’s models, and that Dario, Sam Altman, and Elon Musk all signed onto it.
Sacks was blunt in his rejection: “It’s not an SRO because it’s not self-regulating. It’s a regulatory organization, and the S part is a total fig leaf designed to hide the fact that this is a new regulatory agency that reports to the government that does pre-release testing and approval of models. I call it a DMV for AI.”
His preferred model is the MPAA — a genuinely self-regulatory organization that does not report to the government, was formed by the motion picture industry in response to the Hays Code, and forestalled government intervention by promoting its own standards. Sacks claimed to have discussed this with Elon Musk, who “does not support a FINRA for AI” but “could get behind an MPAA for AI.”
Chamath countered that the National Futures Association might be a better SRO model given its oversight of a more dynamic market. Sacks dismissed the distinction: any government-reporting body with pre-release testing authority becomes a queue for model approval, slowing the industry while China accelerates.
He cited China’s successful steel rocket landing as evidence of the regulatory gap: “Elon’s still waiting for FAA approval on his steel Starship.”

The hosts agreed that regulatory capture will eventually target open-source models. Sacks laid out the scenario with precision — and predicted it will be framed as fairness.
“An open source ban is coming. They’re not going to call it that. You know what they’re going to say? They’re going to say that we simply have to apply the same standards to open models that we apply to closed,” Sacks said.
The logic chain is straightforward:
Sacks cited Dario’s 2023 Senate testimony, where he argued that open models are dangerous precisely because they “cannot be centrally monitored and controlled and rolled back and withdrawn” — characteristics that are technologically immutable. The hosts agreed this logic will be used to justify open-source restrictions under the guise of uniform standards.
Friedberg offered a crucial technical rebuttal. “When you take a closed source model, it does really well on a benchmark. When you wrap it in its own harness, it decays in capability. When you take an open source model and you use any other open source model, it improves in capability,” he said.
His argument: open-source models are not just philosophically preferable — they are becoming more performant and cheaper. “Unless America stops us from using the cheapest, fastest, best thing, we will win.”

Chamath introduced a deeper philosophical challenge: if recursive self-improvement is real — the theory that AI models can spin up agents to build better models without human intervention — then all regulatory schemes are “a fool’s errand.” RSI doesn’t require government approval, human architects, or even a specific geography.
“All that it needs is the chips and the power and the connection. And if it’s got the chips and the power and the communication connection, it can run… the labs can go and set up data centers in Iceland, in China, in wherever. Whoever will give them the freedom,” Chamath said.
His argument for keeping frontier labs in the US: “Wouldn’t we be better off having systems where we actually have those labs based, where we have those data centers based in the US, where we have jurisdiction over them, where we have capacity to monitor and control and police them rather than just have them sprung up in some random country or some random territory?”
Friedberg raised a critical transparency objection: frontier labs obfuscate their “thinking tokens” — the intermediate reasoning between input and output. If these companies genuinely feared misalignment, they would un-obfuscate these tokens so third parties could verify their claims. Instead, he argued, “we have to basically agree to their interpretation of tokens that we can’t see, that only they can see, that they can interpret how they want.”
Sacks’s response to the RSI scenario: “I don’t think some new government regulatory agency will stop it if that’s the thing because it won’t be technically sophisticated enough. All it will do actually is concentrate power around that thing.” He argued for maintaining choice — multiple AGIs with different alignment approaches, whether Elon’s “maximally truth-seeking” or Anthropic’s “cloud constitution” of liberal values — rather than a single regulated path.
Sacks was careful to distinguish his position from anti-safety sentiment: “Just because I believe that we should not have a DMV for AI does not mean that I’m against AI safety.” The relevant pressure, he argued, is already being applied by the courts. He cited Meta’s product liability exposure — over $1 trillion in pending lawsuits, including a recent $6 million jury award in Los Angeles for creating body image issues. AI labs, he predicted, will face similar legal reckoning if they race products out without adequate safety measures.
Friedberg warned that over-regulation would not just slow innovation — it would trigger a capital exodus from the United States.
“What you will see is FDI, foreign direct investment in the United States go off of a cliff. And what you will see is direct investments into other countries of the world rise,” Friedberg said.
He argued that multinational companies like Coca-Cola and Pfizer would shift investment and operations abroad if the regulatory burden becomes too heavy, similar to what happened in China after the crackdown on tech companies. The FDA’s long and slow approval cycle was cited by Chamath as a reason biotech investment has already moved offshore to China.
Friedberg also raised a technical objection to the transparency demands embedded in the SRO proposal: “If all of a sudden competitor A, B, and C could see every single trace and span that your model generated as an artifact of going from input tokens to output tokens, you’re essentially giving away a lot of your recipe.” The proposal, he argued, is technically impractical.
The regulatory debate is unfolding against the backdrop of explosive commercial success for the very companies being discussed.
Anthropic just reported preliminary second-quarter revenue exceeding $11.5 billion, up from $787 million in the same period of 2025 — a 14-fold increase. The company posted its first-ever operating profit since inception. Its annualized revenue run rate crossed $65 billion by the end of July, according to Bloomberg.
OpenAI, meanwhile, generated $6.7 billion in revenue in the second quarter, with an annualized run rate exceeding $40 billion. Both companies are racing toward IPOs: Anthropic is preparing to publicly file paperwork as early as the end of this month, while OpenAI has told employees it expects to be public in 2027.
The commercial stakes make the regulatory positioning more consequential. Anthropic is reportedly seeking a valuation around $1 trillion, anchored to a 2028 revenue projection of $190 billion to $200 billion. If public investors accept that framework, it would establish a template for the entire AI industry.
Sacks drew a direct line between Meta’s legal exposure and the AI industry’s future. Meta faces over $1 trillion in product liability lawsuits, including a recent $6 million jury verdict for creating body image issues. Sacks noted that OpenAI is already slowing frontier model releases due to safety stack concerns — driven, he argued, by this liability exposure.
His prediction: AI labs like OpenAI and Anthropic will face massive product liability lawsuits similar to Meta’s if they race products out without adequate safety measures. The courts, not a new federal agency, are the appropriate check on reckless deployment.
The hosts located the AI backlash within a broader political shift that they found deeply concerning. Chamath presented data from a July 2026 Wall Street Journal article and Fox News poll:
“I don’t think that this is a principled line in the sand anymore,” Chamath said. He argued that unaffordability — driven by government spending in education, healthcare, and housing — is pushing voters toward socialism across both parties. His prediction: a wave in 2026, followed by an Alexandria Ocasio-Cortez presidency in 2028 and a “politburo” featuring Ro Khanna as Secretary of Housing and Urban Development.
The economic picture Chamath painted: the top 50% of Americans hold $178 trillion in net worth versus $6 trillion for the bottom 50%, and 63% of Americans live paycheck to paycheck. The total net worth of all people worth over $50 million is only $23 trillion — taking 10% would cover just three months of current fiscal spending. Sacks cited a Cato Institute estimate that the Democratic Socialists of America platform would cost between $71 trillion and $211 trillion, versus $35 trillion in total corporate profits and $6.6 trillion in Forbes 400 wealth.
Sacks pushed back with a structural analysis: inflation is concentrated in government-heavy sectors. Over 26 years from 2000 to 2026, overall inflation rose 96%, healthcare rose approximately 300%, but computer software fell 67% in real dollars, and average hourly wages grew 135%. “Socialism causes the problem. And then what happens is that groups like the DSA argue that the way to solve the problem is through more socialism. So it’s a disease that claims to be a cure,” Sacks said.
Friedberg offered the sharpest synthesis of the political moment: “This is not about mathematical realism. This is about anger towards a class of people that are increasingly odious and misrepresented.” Corporate America and Silicon Valley, he argued, have “completely lost the script” — no longer a place for “weirdos” and idealists, but a “credentialing place” focused on making money. His prescription: the White House needs to get corporate CEOs in a room and tell them to “play nice in the sandbox, because you’re about to lose the standard operating procedure that allowed you to exist in the first place.”
Chamath offered the episode’s most striking synthesis, connecting the AI regulation debate to the socialist surge through the metaphor of the data center: “The data center is the temple where the billionaires and trillionaires pray. It’s where they go to make their sacrifices. It is their gathering spot. So I think the data center is this like, you know, temple on the mountain. And so that’s why I think the data center gets tied up in all of this. There’s no rational discourse about data centers.”
Friedberg extended this through the Rene Girard lens of scapegoating: “The billionaires will be scapegoated and the trillionaires will be, I mean, even worse. And because that will allow the catharsis that will allow society to reset.”
Calacanis offered a pragmatic prescription for the administration’s second half: focus on affordability with the same executive energy applied to data centers and crypto. His specific asks: a national push to build one to two million homes, trade schools for the “tool belt generation,” and a minimum wage increase. He noted that Meta’s Mark Zuckerberg has been the most effective spokesperson for the industry, creating a trade school program for blue-collar construction jobs, while Dario and Sam have failed to explain how AI will create jobs.
Calacanis shared his personal experience as evidence: his company has seven job openings because teams that embrace AI tools generate more business opportunities, requiring more hires. “If you can convince your team to embrace the tools, so many business opportunities emerge that you wind up hiring people to work for the people in your organization who embrace the tools.”
The hosts could not agree on whether the US can maintain its AI lead while addressing the legitimate economic anxieties driving the backlash. Sacks predicted with high confidence that China will win the AI race if the US adopts a FINRA-style regulatory agency, because China will not subject its companies to such constraints. He dismissed the polling data that suggests Democrats will take the House in November, citing political scientist Patrick Graffini’s analysis of over 3,000 polls across four election cycles since 2018, which found a consistent D+3.7 polling error favoring Democrats.
Friedberg’s final warning was economic: “If all of a sudden competitor A, B, and C could see every single trace and span that your model generated… you’re essentially giving away a lot of your recipe.” The SRO model, he argued, would destroy the competitive advantage that made American AI dominant in the first place.
The unresolved question at the episode’s core is whether Dario Amodei — who argued last year for “data centers full of geniuses” as the key to American dominance — will step forward to defend data centers against the very regulatory apparatus he helped legitimize. The hosts’ answer, implicit throughout the 90-minute discussion, is that he will not — because his regulatory vision and his commercial ambition are, in their view, ultimately the same project.
The stakes, as Sacks framed them in his closing: “America is not gonna be the number one country in the world economically and militarily if we lose AI. It’s just that simple.” The question left hanging is whether the political system can avoid the regulatory overreach that would make that outcome inevitable.
References:
Once added, BigGo Finance appears first in Google Search Top Stories, so you get the broadest, most up-to-the-minute, and most comprehensive global financial news first.
The news and data on this website are for reference only and do not constitute investment advice or an offer to buy or sell. Information is sourced from exchanges and public sources, and may be delayed, interrupted, or updated. While we strive for accuracy, we do not guarantee timeliness, correctness, or completeness. Content may include external links for which we are not responsible. By using this site, you agree that we and our partners are not liable for any losses. Investment carries full responsibility; please carefully assess risks and consult professionals. If there are errors in the content, please contact us for correction.

source