Sacks argues DeepSeek's open-source strategy is about undercutting leading American companies, not altruism.
“if you're if you're behind and you're trying to catch up, then open source is a strategy that actually really makes sense for you. Absolutely. Know, they're trying to basically undercut the leading American companies.”
Sacks argues the $6 million DeepSeek cost claim compares final training run to fully loaded US company costs.
“I mean, if you were to make the apples to apples comparison, you would need to compare the final training run cost by DeepSeek to that of OpenAI or Anthropic.”
Sacks frames AI as the difference between 4% great growth and 2% modest growth for the economy.
“So AI is the difference between having great GDP growth, say around 4%, and modest GDP growth around 2%.”
Sacks contrasts Silicon Valley's ecosystem thinking with Washington bureaucracies' desire for control over exports.
“You want to have lots of participants in your ecosystem. That's a good thing. But that's a hard concept for Washington to wrap its head around sometimes because what Washington wants is control.”
Sacks says October 2023 restrictions assumed US semiconductor monopoly justified any export controls.
“We were the only ones who could really make advanced semiconductor, so therefore, we could impose whatever restrictions we Nobody would have a choice.”
Sacks frames data center decision as binary: countries will choose American or Chinese technology.
“Every sovereign country that can afford them is gonna have data centers. Is that gonna be American technology or Chinese technology?”
Sacks argues DEI ideology was being deliberately programmed into AI models under Biden, mirroring social media censorship apparatus.
“It's because the values, the ideology was being programmed in. Now I think that that was the track we were on before president Trump got elected.”
Sacks claims AI doomerism on the left is designed to justify government intervention and surveillance of AI models.
“Because when you create enough fear in a population, the people will cry out for government to intervene and save them.”
Sacks reports a thousand AI bills are moving through state legislatures with over 100 already passed creating contradictions.
“You've got a thousand bills right now going through state legislatures, over a 100 measures have already passed, and you're starting to see contradictions emerge in the way that the states are regulating this.”
“And here Bernie is acknowledging that this is not about states' rights because he's saying that there needs to be a moratorium on new data centers even if the states want them.”
Sacks argues usage and market share, not other metrics, prove success in AI competition.
“I mean, in this case, the way that you know that you're succeeding is through market shares, through usage.”
Sacks criticizes European regulators for viewing AI leadership as setting regulations rather than driving innovation.
“Now, when I hear European regulators talk about AI, there's a little bit of main character syndrome, where when I've heard some European leaders talk about AI leadership, they're talking about setting the AI regulations and somehow the rest of the world's gonna copy that.”
Sacks argues patchwork state AI regulation burdens startups while big tech can comply, threatening US seamless national market advantage.
“The big tech companies are gonna figure out how to comply with that, but it's a tremendous burden for small tech, for startups, for entrepreneurs.”
Sacks says 1,200 AI bills are moving through state legislatures with 50 states going different directions.
“You've got 1,200 bills right now going through state legislatures to regulate AI. It's a little bit of a knee jerk reaction. And regards you've got 50 states running in 50 different directions.”
“You know, if we look around the world in five years and we see that American companies, American technology has say 80% market share, it means we did a good job.”
Sacks says 1,200 AI bills are currently moving through state legislatures in a knee-jerk reaction.
“There's over 1,200 bills going through state legislatures right now. I think it's very much a knee jerk reaction.”
Sacks says Anthropic wants a permissioning regime requiring Washington approval to release models or sell GPUs anywhere in the world.
“They do want a permissioning regime in Washington for chips and models, meaning you have to go to Washington to get permission to release new models or to sell GPUs anywhere in the world.”
Sacks warns against banning AI for kids when China is incorporating it into K-12 education.
“In China, they're incorporating AI into k through 12 education. Are we gonna ban it for our kids and teenagers? I think that'd be a terrible mistake.”
Sacks says hyperscalers pledged not to raise residential electricity prices in exchange for easier permitting to build own power generation.
“But in exchange for that, the quid pro quo is we'd make it easier to do permitting and make it easier for them to basically stand up their own power generation.”
Sacks cites Yale Budget Lab finding no discernible labor market disruption in first three years after ChatGPT launch.
“Yale Budget Lab did a study and they said that there was no discernible disruption, I mean, that's like a quote, to the labor market in the first three years after the launch of ChatGPT.”
Sacks says $650 billion hyperscaler CapEx this year creates 2% GDP tailwind, excluding energy and productivity gains.
“I think $650,000,000,000 this year just in CapEx from the hyperscalers. That's about a 2% tailwind to GDP just from all the construction that's happening. I don't think that includes essentially all the energy.”
Sacks questions AI CEOs predicting developer obsolescence while Anthropic recruits developers at $500,000 annually.
“I saw that I think Anthropix got numerous open job recs for software developers at $500,000 a year, which seems like a lot to pay someone whose job is going to be obsolete and not needed anymore in just a few months.”
Sacks says less than 5% of layoffs last year were attributable to AI according to Challenger Gray report.
“Less than 5% of the layoffs last year were attributable to AI. It was actually just a very small category.”
Sacks says AI infrastructure boom caused blue collar trade wages to increase 25-30% due to worker shortage.
“Wall Street Journal had a really good article about this. The construction industry is seeing a ton of demand.”
Sacks says Trump proposed letting AI companies build their own power generation behind the meter over a year ago.
“And I think a really good example of that was his idea around data centers where he said over a year ago, before data centers even became a hot political topic, that we should let our AI companies stand up their own power generation behind the meter.”
Sacks warns that government, not individual actors, poses the biggest AI centralization risk for surveillance and censorship.
“I think it's government that's gonna do that, not necessarily an individual actor because it's governments that ultimately have the power.”
Sacks argues DSA candidates are taking over the Democratic Party and the establishment has lost control.
“And you can see the Democratic establishment is in complete panic right now because they have lost control of the party to Zoram Mamdani and his his allies.”
“I've been saying for months that we're on a shot clock. We have to do smart things. We can't just slow everything down because that will not slow down the Chinese.”
Sacks says China is typically nine months behind US AI models but can catch up faster on specific breakthroughs.
“I believe them because, look, the the Chinese have been, I'd say, nine months behind our models, plus or minus three months depending on capability.”
Sacks asserts Chinese GLM model now matches currently available OpenAI and Anthropic models in capability.
“So we now have a Chinese open weight model that is as good as the currently available models from OpenAI and Anthropic.”
Sacks argues Anthropic and OpenAI form emerging duopoly and Anthropic pushes regulatory capture through safety claims.
“As far as we know, no one else is really generating meaningful revenue at the model layer. So we already have, let's say, an emerging duopoly situation.”
Sacks explains export control letter required three conditions including Dario calling model cyber weapon and Amazon reporting failed guardrails.
“Number two is you have a trusted partner of Anthropic, which is Amazon, which did its testing of Fable, which was Mythos with guardrails, and they reported that the guardrails failed.”
Sacks endorses Zuckerberg's framing that the defining question is whether superintelligence access will be centralized or universal
“Mark Zuckerberg gets it right: “The defining question of our age isn’t whether superintelligence will exist, but who will have access to it. Will it be centralized and restricted to a few institutions, or will it be a tool that empowers everyone?” Concentration of power is the https://t.co/kFgkv5dlwF”
Sacks lists five reasons frontier labs call for slowdown: virtue signaling, CYA, regulatory capture, groupthink, and monopoly masking.
“Number one is virtue signaling, and that can never be underestimated as a as a motive in Silicon Valley.”
“Some thoughts on Dario’s post: 1. Dario does not actually address Gavin Baker’s account of what he said – something he could easily deny if it were inaccurate. 2. Dario claims his critics live in a “bubble” where all regulation equals regulatory capture. He calls this an overly https://t.co/Pz3HFfk3Uy”
Sacks argues bio risks can be addressed by regulating DNA synthesis at physical chokepoints rather than creating new regulatory agencies
“@JTLonsdale Bio is an area of concern, but that doesn’t require a new rule-making agency with expansive powers. We can regulate DNA synthesis at the physical chokepoint, as this letter proposes. Targeted solutions for specific risks is better than new bureaucracies. https://t.co/mD5LEovKZK”
Sacks warns that certain AI policies would disadvantage the United States compared to China.
“handicap the U.S. relative to China, which will not adopt the same constraints.”
Sacks claims that Anthropic CEO Dario Amodei's public statements have significantly aided efforts to ban AI data centers in the United States.
“Dario’s messaging has been massively helpful to efforts to ban data centers here in America.”
Sacks argues Trump's requirement that AI companies build their own power generation prevents residential price increases and can lower prices
“President Trump was ahead of the curve requiring AI companies to build their own power generation so new data centers don’t raise electricity prices for residential ratepayers. Done right, data centers actually lower prices by producing excess power and funding grid upgrades.”
Sacks challenges Dario to speak out against over-regulation of data centers if he believes his own argument.
“And probably it would be a good thing if Dario really does believe in the data centers full of geniuses idea that he was arguing for last year, He should come forward and say, we're making a big mistake by over regulating data centers. Here's how we should do it.”
Sacks recounts Dario's argument that having thousands of AI agents versus China's dozens would ensure US dominance.
“And if we have thousands or millions of those and China has only dozens or a 100, whatever, then we're going to win everything.”
Sacks questions whether Dario only fights for doomer arguments and not positive AI policy positions.
“And it makes you kind of question, do you only fight for the doomer side of the argument? Or are there versions of the positive side that you're willing to fight for politically?”
Sacks says restricting allies from data centers means the US is shooting itself in the foot.
“So I guess our allies aren't gonna have the data centers either. So we are totally shooting ourselves in the foot.”
Sacks notes Dario's argument for AI advantage over China contradicts current data center bans.
“And if we have thousands or millions of those and China has only dozens or a 100, whatever, then we're going to win everything. And that was the whole argument.”
Sacks argues current policies ban the very data centers Dario advocated for and restrict chip sales to allies.
“But now we're basically banning the data centers. We're not gonna have those data centers full of geniuses. And he's done his best to restrict sale of chips to our allies as well.”
Sacks says Dario argued for US data centers to create thousands of Nobel Prize-level AI scientist agents.
“Dario actually made this argument in the context of trying to ban export of chips.”
Sacks predicts an open source AI ban disguised as equal standards for open and closed models.
“An open source ban is coming. They're not going to call it that. They're going to say that we simply have to apply the same standards to open models that we apply to close.”
Sacks argues government officials lack technical expertise to set AI safety standards independently.
“Because frankly, people in government have no idea how to monitor and control and set standards for AI safety. Technologically, this is way beyond them.”
Sacks claims Dario is pursuing a FINRA-style body as a Trojan horse for FDA-level AI regulation.
“Dario wants an FDA for AI, but he doesn't have enough political support for that. So instead, they do this Trojan horse of a FINRA for AI.”
Sacks says Anthropic's premium pricing depends entirely on maintaining a six-month lead over open models.
“Because their whole right to charge this massive premium for tokens is because they're six months ahead of those open models.”