AI Cyber Panic, Ken Griffin’s Bubble Warning, The Fall of France?
Chamath PalihapitiyaJason CalacanisDavid SacksDavid FriedbergNikesh Arora
- Chamath sees AI’s bubble risk in concentrated, venture-funded demand rather than in the technology’s usefulness. His estimate puts 90% of AI spending in 1% of companies, with 80% of that coming from unprofitable VC-backed businesses. He describes VC-funded startups circulating spending, showing revenue traction, then raising again from the same investors. “AI has a diffusion problem”; Jason counters that startups’ rush toward the cheapest, fastest competitive advantage is an encouraging adoption signal.
- Open models could redirect AI spending toward infrastructure, but compute scarcity limits enterprise independence. Nikesh Arora says Palo Alto Networks prefers open weights for customization and auditability, yet cannot easily procure globally resilient inference outside hyperscalers. His likely destination is rented bare metal running open models: buyers remain “a taker of terms,” while profit pools might shift from chips and memory toward available compute.
- Arora expects AI-discovered vulnerabilities to arrive faster than enterprises can clear their accumulated technical debt. Jason describes PHP exploits rated 9.2/10 affecting software used across roughly 70% of websites; Arora explains how persistent compute unlocked an attack previously considered impractical. More defensive compute alone cannot fix fragmented infrastructure, making the operational priority “how quickly can I find them and how quickly can I eject them?”
- Banking agents expose an unresolved boundary between user authorization and defective software. Arora and Chamath put responsibility primarily on users who hand over credentials, while Friedberg sees potential provider liability when an agent disregards bank access restrictions. Arora preserves a crucial exception: a “rogue agent” that fails to behave as advertised creates a product-liability issue.
- France becomes the episode’s sovereign-risk warning, with disagreement over the relative roles of government provision and immigration. Jason reports French 10-year yields above 5%; Chamath ultimately calls for bond vigilantes to push them toward 8% and force a reset. Chamath warns that political actors can capture power during a reset; Jason, doubting AI abundance would be persuasive to France, concludes, “I don’t see a reasonable endgame.”
- Friedberg’s fiscal proposals include winding down federal student lending, changing housing finance, and transitioning defined-benefit retirement promises into individually owned investments over a decade. David Sacks supplies the more specific proposal to reduce federal home-loan underwriting, warning of a possible 10–20% hit to middle-class net worth. Chamath supports making student debt dischargeable; the retirement transition would still remove guaranteed income.
- The closing dispute is whether AI-led growth reaches voters without more direct measures on affordability. David Friedberg calls the US economy “entirely and completely levered to and correlated with the AI trade,” while Jason wants equal attention to housing and lower-paid workers. Jason proposes $1 annual minimum-wage increases toward $15; Friedberg argues for tax relief and credits instead, rejecting government intervention in labor prices.
1. AI’s bubble question is about who pays the bills
Jason introduces Ken Griffin’s warning that AI is “almost certainly a bubble,” alongside Griffin’s longer-term optimism: major economic transformations commonly bring bubbles. The immediate concern is businesses examining Anthropic and OpenAI bills and asking where productivity appears. Jason reports Thursday declines of 10% for CoreWeave, 6% for Meta, and 5% for AMD and SpaceX, while noting recent volatility and a generally upward trend.
Chamath’s estimate puts 90% of AI spending in 1% of companies, with 80% of that spending coming from unprofitable VC-backed businesses. His causal chain runs from LP capital to venture funds, startups, token purchases, revenue growth, and subsequent fundraising. The danger is “a bunch of money-losing entities just spinning a bunch of money around,” supporting valuations before demand broadens.
An unpublished Ramp chart and a call with its chief economist broadly support Chamath’s estimate, according to his researcher. Jason flags Ramp’s concentration in startup financial services; Chamath says its economist also accesses broader data. Chamath contrasts today’s circularity with early AWS adoption: startups initially dominated EC2 and S3 usage, but Amazon’s retail business was roughly break-even on free cash flow.
Jason’s pushback — worth keeping: startups adopting a technology aggressively can reveal genuine economic advantage. They seek the “cheapest, best, fastest edge,” unlike slower enterprises. His examples include cloud computing and WeWork as cost-saving advantages, and initially underpriced Facebook advertising, into which startups poured venture dollars because it worked. Concentrated early adoption therefore can be a leading indicator of diffusion rather than evidence that demand lacks substance.
2. Open weights attract enterprises, but infrastructure remains scarce
Chamath reads rising GPU-compute spending in Ramp’s chart as evidence that heavy users are moving toward open models. He also reports changed enterprise buying patterns following Alex Karp’s comments about data sovereignty. One unnamed major US bank told his team it had reduced AWS and GCP spending and built six independent, ground-up data centers: “It’s back to the future.”
Arora says Palo Alto Networks prefers open-source, open-weight models because it can modify weights, train them, and interrogate their chain of thought. His enterprise requirements are “explainability, auditability, and verifiability.” In regulated financial services or healthcare, telling a boss that an unexplained result came from an AI model will not suffice; he sees open weights as necessary for that scrutiny.
Independence becomes harder at production scale. Arora needs global inference with 99.9% reliability, not a startup’s single 10-megawatt deployment that can tolerate several hours offline. He says Jensen cannot supply chips until 2028, while land, power, and buildings also constrain supply. Having closed Palo Alto’s data centers six or eight years earlier, he now finds rebuilding them an impractical negotiating threat.
Arora expects a hybrid outcome: hyperscalers or adjacent providers supply bare metal while enterprises run open models. The “profit pool” could move from NVIDIA and memory suppliers toward owners of available compute. He acknowledges slow enterprise diffusion but calls demand “insatiable”: even if every company committed immediately, he argues, there would not be enough compute to serve them all.
3. The PHP panic turns neglected maintenance into an emergency
Jason describes two critical exploits nicknamed “PHP2Shell,” discovered using AI in August and rated 9.2/10. He says PHP powers about 70% of websites, including WordPress and Wikipedia. Developers had coordinated an October 22 patch, but information leaked beforehand; PHP advanced the release to October 21. How the information escaped remained unknown during the recording.
The release coincided with increased hacking activity, according to Jason’s account of AWS and Cloudflare reports. He describes millions of attempts, precautionary shutdowns across European government websites, systems taken offline at Wikipedia, and NSA/FBI warnings. His question is whether such emergencies will recur every ten days. Arora’s answer is sharper: “I think 10 days is too long.”
Arora’s concrete example is a CIO whose websites run different PHP generations. Current versions can accept the patch; older ones require taking sites down and rebuilding them. The organization had tolerated those older versions because upgrading never seemed important enough. AI has changed the urgency of paying technical debt, rather than creating all the underlying weaknesses.
There are two defenses in Arora’s account: patch the vulnerable application itself, or deploy a “virtual patch” through a perimeter product such as Palo Alto’s or Cloudflare’s. The latter blocks the recognized attack before it reaches the server. But publication also informs attackers: they inspect what changed, reverse-engineer the vulnerability, and target organizations that have not yet patched.
4. Defensive compute needs coordinated systems to become useful
Arora says powerful models can find vulnerabilities and chain several into an attack path when given sufficient compute and persistence. The PHP problem, he explains, had been identified years earlier but seemed too difficult to exploit because of the coordination required. Persistent computation found a workable path. Once disclosed or reverse-engineered, that discovery became usable by many more attackers.
Jason asks whether well-funded defenders can simply outcompute hackers. Arora says that is not true today because infrastructure lacks the necessary orchestration. A company with 5,000 websites may still have IT staff cataloging them individually rather than operating from a unified inventory. Without that baseline preparedness, defenders cannot dispatch coordinated agents to patch systems merely because more compute is available.
Continuous penetration testing helps, Arora says, but each stronger model may discover additional entry points. His alternative starting assumption is that breaches will happen: “The question becomes less about how I got breached” and more about detection and ejection. Aggregating signals and data makes anomalous behavior visible, allowing superior defensive compute to operate against an intruder rather than an unknown inventory of vulnerabilities.
5. Banking agents blur identity and divide liability
Friedberg separates suspicious browser activity from conventional application access through verified APIs and intermediaries such as Plaid. Arora likewise distinguishes two problems: unauthorized access possibly connected to the PHP vulnerability, and personal agents logging in with their users’ credentials. Banks cannot readily distinguish the latter from users themselves. CAPTCHA is an incomplete boundary because AI can solve it or hand the challenge back to a person.
Jason asks whether banks are exaggerating danger to prevent agents from routing customers toward better rates and cheaper services. Arora considers that too much strategic credit. His explanation is liability and support costs: a customer says $50,000 disappeared, while the bank must reconstruct actions performed by software holding valid credentials. Repeated requests for explanations could become expensive even before reimbursing disputed transactions.
Friedberg’s initial liability view places responsibility on an agent provider if its software ignores a bank’s prohibition on agent access. If the software warns the user and receives an explicit override, he thinks responsibility shifts to the user. Jason supplies an e-commerce parallel: his Grok shopping agent can prepare an Amazon cart, address, and payment choice, but “I won’t hit the buy button.”
Arora’s default is user responsibility: “There’s a word in agent, it’s called agency.” Chamath calls that the only scalable arrangement. Jason pushes back on agents that find unauthorized workarounds to satisfy instructions, prompting Arora’s exception: a “rogue agent” that behaves contrary to its advertised purpose creates a product-liability problem for its maker. Their agreement on authorization does not erase that distinction.
6. Removing model guardrails produces the sharpest AI disagreement
Jason describes incubated company Obliteration AI as supplying models without built-in constraints for white hats and corporations, with customers responsible for controls and liability. He frames the product as infrastructure for serious testing rather than unrestricted public distribution. Arora asks what this adds beyond taking an open-weight model and training it internally; Jason’s answer centers on serving infrastructure and efficiency.
Chamath finds this “unbelievably scary,” calling it an underappreciated risk vector that concerns him more than alignment. Arora argues governance depends on the weakest link: restrictions elsewhere accomplish little if another provider supplies raw capability without guardrails. His analogy is handing a nuclear bomb to anybody who can afford it. Jason maintains that qualified defenders also need powerful tools to strengthen their systems.
Arora separately rejects Jason’s theory that frontier labs cause the security problem by withholding their strongest defensive capabilities. He sees researchers struggling to package powerful systems under compute constraints; even his access to recent publicly available models remains rate-limited. Their commercial motive is more basic than a deliberate attack on cybersecurity incumbents: “They just need money, bro,” including money ultimately owed to Jensen.
7. Security spending and management incentives have to change together
Chamath explains why technical debt persists: engineers prefer the next product to maintenance, while privately deployed software drifts into many versions with different upgrade schedules. Arora adds the incentive problem: CIOs are rewarded for keeping costs low and systems available. Upgrades cost money, introduce downtime risk, and generate complaints, so the behavior that preserves today’s service can accumulate tomorrow’s security exposure.
Arora puts cybersecurity spending at roughly 4–10% of IT budgets. Cloud-oriented businesses can sit around 4–5% because hyperscalers handle infrastructure security; financial services and the federal government are nearer 8–10%. He says customer calls increasingly ask how to stop getting “punched in the face every morning.” Jason expects the transition to raise both security’s importance and its claim on scarce AI talent.
Beyond existing defenses, Arora says enterprises still need security systems for AI itself: discovering agents, inspecting their actions, and controlling information leakage. He cites a researcher finding that a shared Muse conversation could expose private information. Yet he also sees opportunity: as finance, HR, and marketing all seek AI expertise, “the CIOs could become the next CEOs.”
Asked how he would respond to Bernie Sanders demanding an AI stop, Arora answers, “Why don’t you start? I’m right behind you.” He argues AI’s builders have overemphasized alarming capabilities and created a “race to the bottom” in public messaging. Jason sees the political consequence: predictions of lost jobs and catastrophic risk give populists on both the left and right an attractive campaign issue.
8. France’s unrest is becoming a financing problem
Jason reports more than half a million participants, damage to over 3,000 businesses, more than 1,000 injured police, and 2,500 arrests. He describes protests originating in dilapidated schools before becoming broader unrest and looting, including destruction of Limoges’ 143-year-old city hall. His account also includes spillovers into Italy, Spain, and Albania, along with French curfews and conflicting government explanations of the violence.
The market figures sharpen the discussion: Jason says France’s 10-year yield exceeded 5%, its highest since 2002, while the US equivalent slipped below 5%. He cites a possible flight from European into US debt. Polymarket prices mentioned on the show imply an 18% chance of a French emergency declaration by year-end and 91% odds of a Moody’s or S&P downgrade by October 31.
Chamath condemns the violence as domestic terrorism and argues the government created a “tinderbox” by failing to confront it. Economically, France’s proposed €50 billion austerity package would need to grow, he says, making additional education investment harder even if school conditions were the central grievance. He predicts the government will lose the coming presidential election and face a complicated bureaucratic and economic reset.
9. The hosts disagree over what France’s social contract proves
Friedberg resists Jason’s comparison with BLM and January 6, which he sees as tied to more specific political events. France’s complaints seem broader: schools, healthcare, jobs, and Palestine. His diagnosis is a cycle in which citizens demand more state provision, services disappoint, and the disappointment creates demands for still more spending. “The more the government tries to provide to the people, the more the people are let down.”
Friedberg calculates US public spending at $12 trillion against $33 trillion of GDP, or roughly 36%, including $7.4 trillion federal and $4.6 trillion state/local spending. He puts both countries near 120% debt/GDP, with deficits of 5.4% in France and 6.2% in the US. Growth of roughly 0.5% versus 2.3%, he argues, explains why France could reach the fiscal cliff first.
Chamath’s challenge is that France and Italy already provide roughly free healthcare and education, yet still face unrest. Jason cites French university fees around €300 annually against approximately €12,000 in state cost. Friedberg disputes that low prices mean the problems are solved: weeks-long medical waits and classrooms of 60 students can leave people dissatisfied despite formally available services.
A later chart puts French tax revenue at 46% of GDP versus 22% for the US. Jason challenges its persuasiveness: many Americans want the healthcare and education associated with countries above the US on that list. Chamath rejects the claim that most want to live there; Friedberg argues the comparison must include employment and advancement alongside public services.
10. Proposed affordability reforms would redistribute gains and losses
On education, Friedberg would end the federal student-loan program while creating a way to discharge existing debt, potentially through bankruptcy. Chamath is “extremely sympathetic” to cancellation and says the debt should always have been dischargeable. Friedberg insists relief must accompany termination of new federal lending, otherwise the price-support mechanism remains. He expects institutions and tuition to adjust over roughly ten years.
On housing, Friedberg argues for changing housing loans that inflate home costs. David Sacks later makes the more specific proposal to reduce underwriting in the federal home-loan system, warning that this could cut middle-class net worth by 10–20%. His premise is that financing has supported existing homeowners’ asset values while excluding younger buyers. Maintaining those values preserves current owners’ solvency and sense of wealth, but broader access requires existing middle-class, upper-class, and wealthy owners to accept a hit.
Healthcare is Friedberg’s least complete proposal: “Healthcare’s the hardest.” He would put more responsibility on employers purchasing insurance, expecting competition to pressure prices and services. He names HIPAA and state/local housing regulation as barriers in their respective markets. These are multi-year changes in his framing, not immediate fixes; healthcare’s complexity remains explicit even as he argues for reducing government intervention.
11. Retirement reform would replace guarantees with ownership
Friedberg proposes making defined-benefit retirement programs, including Social Security, illegal and moving them into defined-contribution accounts over a decade. Each year, 10% of an individual’s promised amount would transfer into an IRA-like or Trump account, with investment allocations chosen according to age and preference. He acknowledges the objection: people lose guaranteed income. Jason points to Australian superannuation, and David Sacks adds Norway as an example.
Jason connects the ownership thesis to Elon Musk’s stated plan to give every American child one SpaceX share, then worth roughly $150–170, describing the gift as about $10 billion. He imagines wider participation by wealthy donors and separate accounts for education and retirement. The important feature in Jason’s argument is holding productive assets over time: “forcing people to hold onto it makes them part of capitalism.”
12. Immigration arguments split economic selection from cultural assimilation
Jason places much of the blame on immigration and failed integration, while also assigning responsibility to native French society for not integrating newcomers. In the subsequent exchange, a speaker cites annual immigration of 1.5% of the population and argues that inflows should be weighed against growth, employment, and needed skills. Another participant argues for selecting immigrants who can support themselves and share civic principles; the transcript’s speaker labels for these comments are unclear.
The latter speaker distinguishes shared civic values from cultural assimilation, naming women’s equality, individual rights, respect for others’ rights, and civil law taking precedence over religious authority. The discussion presents this as a proposed screening criterion, not a claim that ethnicity or first language should determine eligibility.
Jason says there were reports that Chinese and Jewish businesses were targeted while Muslim businesses were avoided. Friedberg presents an interview expressing resentment that Chinese immigrants had become prosperous. Jason—not Chamath—says the targeting reminded him of a “bottom-up” version of Kristallnacht. This is his interpretation of alleged ethnic targeting and envy, rather than an established explanation for all the unrest.
Friedberg’s most personal example comes from his family’s restaurant: his nurse mother made desserts to save money, and he and his brothers polished silverware and worked from childhood while the family remained financially stretched. That experience underlies his anger at small businesses being destroyed. Both he and Jason argue owners should be able to defend their property with firearms, invoking Korean shopkeepers during the LA riots.
13. Bond vigilantes offer a reset that Chamath also fears
Chamath argues bond investors can force a structural bargain by making French debt unfinanceable. When Friedberg asks whether 6–7% yields would suffice, Chamath answers, “I think you need to take France to, like, 8% and just break them.” His desired result is to put pensions, immigration, service quality, and other commitments on the negotiating table before the underlying social tensions compound further.
Yet Chamath also describes why forced resets fail: political actors accrue power, assign blame, and preserve influence rather than repair institutions. “In the middle of a reset, what happens is you just give away all your rights.” AI abundance might offer another route, but Jason doubts distressed French citizens would trust promises from Silicon Valley oligarchs. His admission remains unresolved: “I don’t see a reasonable endgame.”
David Sacks distinguishes violent social collapse from debt-driven economic restructuring, citing Argentina, Brazil, and Greece. The latter still entails protests, inflation, and lost wealth, but might avoid a deeper social breakdown. Sacks argues US leaders should reduce annual federal spending from roughly $7.5 trillion toward $4 trillion “in a way that doesn’t destroy the economy,” using deliberate reforms rather than waiting for a financing crisis.
14. AI-led growth and voter reassurance are different political tasks
Jason turns France’s discontent into a warning about US political communication ahead of November 3. Trump, he argues, should discuss housing as frequently as data centers and visibly address lower-paid workers alongside technology and large businesses. He credits the administration’s economic activity but says voters hear attention to crypto and infrastructure while their own housing, wages, and healthcare remain unresolved.
David Friedberg’s defense is that the AI buildout carries an outsized share of economic growth and benefits workers beyond the largest technology companies. He points to skilled-labor wages, investment in local communities, and pressure to build data centers behind the meter while lowering electricity costs. His categorical framing: “The US economy is entirely and completely levered to and correlated with the AI trade.”
Jason accepts the growth argument but maintains that its political transmission is weak. He wants explicit commitments to affordable housing and the bottom half of the income distribution, not reassurance that a “golden age” has arrived. The disagreement concerns whether the administration has connected those benefits convincingly to everyday life, alongside the substantive question of what additional policies workers need.
15. Minimum-wage disagreement survives agreement on helping workers
Jason proposes adding $1 annually to the federal minimum wage for seven years toward $15, a destination he dates to 2032. He estimates the change would reach roughly 10–15 million people and argues their extra income would be spent back into businesses. His case combines spending velocity with social cohesion: the president should visibly ask employers to do better for their lowest-paid workers.
Friedberg invokes California’s AB 1228 and its $20 fast-food wage, citing 18,000 lost jobs. He explains the mechanism through a consumer who can afford a $4 burger but not an $8 one: employers cannot indefinitely pay wages unsupported by customer demand. In his account, closures hurt both willing workers and consumers, and even smaller mandated increases can push marginal transactions beyond viability.
Jason rejects the comparison with an abrupt move to $20, says the California policy added about 25 cents per meal, and argues displaced workers found more productive jobs without changing the overall unemployment rate. He challenges Friedberg to reconcile that objection with their usual argument that automation releases labor for better work. Friedberg maintains that Jason is still trying to determine the market’s correct price.
David Sacks’s alternative is lower taxes, credits, and other incentives that increase disposable income without setting labor prices. Jason endorses a zero tax rate up to $150,000 and urges presidential support. That agreement does not settle the original dispute: Sacks says intervention has adverse consequences, while Jason closes by noting the country already has a minimum wage, so the question remains what its level should be.
Full transcript
All right, everybody. Welcome to the literal number one podcast in the world three weeks running. We got a lot to get to. Sax is out this week. We've got a special guest coming up. Let's get to our first story. Ken Griffin says AI is almost certainly a bubble, and Chamath has responded, moving markets almost every week here on the pod. Look out, France. The Citadel founder made the comments this week at the company’s annual Future of Global Markets conference. But he says, generally, “Hey, bubbles are normal. Any radical transformation of our economy is generally intertwined with a bubble.” He said he’s bullish on AI’s long-term potential, but his concern is that companies are struggling to turn AI spending into productivity, something we talked about here last year. Companies are, quote, “looking at their bills from Anthropic and OpenAI and asking where the productivity is.” We’ve had this discussion over and over again.
On Thursday, tech continued to get crushed on the data center and CapEx front. Companies either making data centers, like CoreWeave, were down 10%, or having huge AI spending, like Meta, which was down 6% after earnings. AMD was down 5%. SpaceX was down 5%. But these things have been swinging up and down a bit and generally trending up. Chamath, what’s your take on market sentiment on October 22, 2026? What are your thoughts on how markets are viewing the AI “bubble”? Do you have concerns? Jensen Huang from NVIDIA thinks money is available, and money is going to want this opportunity to finance it. I think we all think companies are getting a ton of value out of AI technologies.
My Social Capital research team had a really great call with Ramp’s chief economist, Ara Karazian. The first question they had was because I was doing my own math, and I tweeted this thing, Nick, maybe you can put it back up there, where I said, “Okay, my back-of-the-envelope math, knowing everything that I know and everything that I see, is about 1% of all of the companies represent 90% of all the revenue.”
Revenue from AI spending?
Yeah. What I said was 1% of companies represent 90% of AI spending—
Got it.
—so that’s tokens. And of that 90%, 80% are from unprofitable, VC-backed companies. The cycle here, just to be very precise, is that VCs raise from pension funds, LPs, foundations, and family offices. They put money into a startup. That startup takes that money and invests in growth. Part of growing is consuming other people’s AI. Then that company starts to show revenue traction. Those folks then say, “Hey, look how well we’re doing,” and raise a bunch of money from the same VCs. It’s just this circularity.
Why that’s important is that what Ken is pointing out is that for AI to really scale as a predictable driver of long-term earnings, it has to diffuse out of the startup ecosystem. You can’t be in a situation where, on the one hand, you’re talking about breathtaking valuations, but 75 cents of every dollar is just a startup sending it back and forth so they can mark themselves up. I think that’s a structural problem.
We’ve seen that structural problem in the past. It happened in the early days of cloud, when AWS initially started to expose EC2 and S3. The consumers of an EC2 instance or an S3 bucket were startups, but the difference there is that AWS was abstracted from Amazon retail, which was already a roughly break-even profitable business, at least on a free-cash-flow basis. What’s different here is that you have a bunch of money-losing entities just spinning a bunch of money around, and it has to diffuse more broadly. I think that’s what Ken is observing.
Yeah.
Now, when we brought that data to Ramp, the quote is—
Oh.
—“incredible,” because this is what my researcher told me. “I prepared a list of factors to see if your thesis”—this is him texting me—“around VC money being 80 to 90% of the top 1% of AI spend,” but he said, “Chamath’s actual estimate is pretty much correct.”
The second thing he sent us—and Ramp hasn’t published this chart, so thanks to Ramp and thanks to Ara—is that he said we could be the first ones to publish this.
And just to be clear for the audience, Ramp provides financial services, primarily neobank credit cards to startups. So they know startup data. Yeah.
I think they know broad-based data from other data sources as well, and I think—
Sure.
—Ara has access to all of it.
But they’re hot here, yeah.
But what he said is: now look at this chart. What this shows is how the top 1% of spenders in AI are spending their money. I think Ramp is going to publish this soon, as I said, so thanks to them for letting us preview this. What’s amazing to me is the blue box in this histogram. It shows what percentage of money they’re spending—
Hmm.
—on GPU compute. Why does that matter? The rose-colored box is LLM traffic. That’s the typical spend you’re going to spend on the big frontier labs, closed models effectively, either through things like Bedrock on AWS or what have you. What’s really interesting is that the blue box is ticking up. The blue box is people buying their own GPUs because they’re moving to open source.
What I think is really interesting is that you have this two-pronged thing happening. One is that the diffusion we need within AI is happening a little bit slower than we would all like. I see that at 80/90 as well. I think the large companies that we work with—you can almost pinpoint it on a chart. Ever since Alex Karp’s commentary on CNBC about data sovereignty, buying patterns have changed, and I think people are being a lot more methodical. They’re building out their own data centers or at least moving into VPCs, virtual public clouds.
My team was at the CIO event yesterday of, let’s say, one of the largest banks in America. You know what they said? I’m not going to say the name of the bank. It was the most incredible thing. They said, “We have decreased our spend on AWS and GCP. We now have 6 completely independent, ground-up data centers that we’ve built from scratch for ourselves.” And I was like, “What is going on here?”
Wow, we’re back to the future.
It’s back to the future.
People standing up their own data centers, yeah.
It’s back to the future. I didn’t mean to have any commentary other than just observing what I think, but AI has a diffusion problem.
Right.
It is too concentrated and ripping around in a small little ecosystem. We need to diffuse out. But as it diffuses out, people are picking open source, is what I would tell you.
Love it. We were early on that discussion two years ago here on the All-In pod. I will say Ken Griffin—and I didn’t hear the full quotes here because I wasn’t in the room where it happened—but based on the quote, I think he may have something slightly wrong.
When startups flood to a technology, that’s a strong signal. Why? Startups—and I invest in 100 a year, and we evaluate over 10,000 a year at my fund—look for the cheapest, best, fastest edge they can get. What was the biggest edge they could get 15 or 20 years ago? It was cloud computing. They saved a lot of money. It was WeWork. They saved a lot of money; it was more efficient. Then it was Facebook ads. Facebook ads were mispriced. They were super cheap. They dumped all their VC money into Facebook ads. Now it’s AI. For the last two years, yes, they’ve been dumping money into it.
If you see startups rush to a technology or a trend and start spending all their VC money on it, it means this little colony of 2- to 200-person companies has determined—and they’re the most nimble—that this is the best solution. So that’s a tell that things are going in the right direction. The big, slow, cumbersome enterprises are the slowest to adopt new technology, and that’s why they buy startups all the time.
Speaking of buying startups all the time, my guy Nikesh is here from Palo Alto Networks to tell us if this is the end of days, if all of our bank accounts and Bitcoin are going to get hacked this weekend, and whether we should buy more guns and chickens and eggs and drill a well. Welcome back to the program, Nikesh Arora. How are you, sir?
I’m doing great. Thank you for having me, guys.
And you had a nice event last week, didn’t you? A nice event in Napa.
I did. We had a bunch of CIOs, not just one bank CIO like Chamath said. But before we jump into the end of days, I do want to say something about what you guys were talking about earlier—
Yeah, please.
—about open source versus hyperscalers providing inference. We had this peculiar problem. We’d like to embrace AI as quickly as many of the startups, and we’ve gone through all the understanding, all the analysis, open source versus cloud models. We prefer open-source, open-weight models because they allow us to modify the weights and train them.
1. Enterprises Turn to Open Models
Something most people don’t talk about is that it allows us to interrogate the chain of thought. Over time, all of these enterprises are going to need explainability, auditability, and verifiability.
100%. 100%.
If I can’t validate it, it’s not going to be okay to go to your boss and say, “I don’t know what happened, but the IT guys put an AI model in there.”
It did it. That’s not explainable to the regulator in financial services and healthcare. They want explainability, verifiability, and auditability. We can only get that with open-source, open-weight models. We can interrogate the chain of thought. That’s one reason we’re gravitating toward them, as opposed to getting caught in the closed race.
The challenge, though, as I think you guys talked about—and I saw Chamath’s chart—is that it’s very hard to find a global, resilient, highly available inference provider outside the hyperscaler community right now.
Right.
So if I’m a startup, great. I can go rent 10 megawatts in one data center in one part of the U.S. If it goes down for a few hours, I’m out of luck, but I’m a startup, so I’m okay.
Right.
But if I’m providing resilient services at 99.9% reliability, it’s very hard for me to find a global inference provider who doesn’t want to sell me their model at a price that makes it economically useful to use open source.
Let me ask you a question, though. Do you think there’s a world where Palo Alto Networks says one of two things? One is you go to the hyperscalers and say, “Okay, listen, guys. You run it for me, but I want the bare metal. I want it totally air-gapped. It’s an AWS facility, but I’ll just lease it from you. Otherwise, just get out of the way so that I can have everything inside of it.” Or do you do what this one big G-SIB bank does and say, “I’m gonna build my own inside the United States”?
Yeah.
Is there a world where you think other companies go to that limit, or do you think that’s an extreme case that’s probably not representative of the norm?
So this is the interesting debate, Chamath. Six or eight years ago, when I came to Palo Alto, I shut down all data centers. I said, “We cannot run a global data center footprint at the scale, speed, and reliability that I could probably rent from hyperscalers.” Life was great, and suddenly what’s happened is that, in the current incarnation of where we are, compute is highly dear. It’s very hard to find compute in the world, period.
So I’m a buyer. I’m a taker of terms, not a dictator of terms. It’s very hard to go and dictate terms anywhere because, sure, you can make any plans you want, but those plans are… You talk to Jensen, he can’t give you chips till 2028. You talk to—where do you get land, power, and shells? So it’s very hard to say, “Yes, I can go build my own data centers,” but it’s kind of an empty threat. It’s very hard to go build them and make them resilient.
And by the way, I don’t run 50 data centers around the world where I’m running 10 to 20 megawatts of inference. It’s a good conceptual idea that I should go become independent and build my own inference. But most likely, the answer is somewhere in the middle. I’m gonna have to get bare metal, yes. Most likely I’m gonna get it from either a hyperscaler or a cousin of a hyperscaler, or somebody’s gonna be stitching together sleeves of compute from around the world to meet my needs, in the hope that I’m gonna migrate to something more resilient and more robust over time.
All right. So, Nikesh—
Right, so there you have a hybrid solution down there. One second. I just wanna complete the thought. You have a hybrid solution down there.
Yeah, I just wanna get to the cyber stuff before we run out of time, just so you know.
Yeah. Then you go and you basically get a bunch of open-source models so that you can interrogate chain-of-thought tokens. Is that a representative solution for most of the Fortune 500 or Global 2000?
I think we get there. I expect at some point in time we will start seeing hyperscalers giving bare metal and letting us run open-source models. They already do to some extent, so that’s probably what’s gonna happen. It’s a matter of the profit pool, right? It all depends on where you shift the profit pool.
Today they all live in NVIDIA and the memory industry, and they’re being shifted toward the hyperscalers, people who have large compute available. At one point, it’ll become less about the chips and more about the availability of compute. I think demand is insatiable. There’s enough demand. There’s more demand than you and I can imagine.
Right. Right.
I understand the enterprise diffusion comment, but it’s gonna get diffused.
Right.
The question is, even if every company in the world today said, “I’m in. I’m in. I’m ready. I’m embracing AI. Let’s go as fast as we can,” there’s not enough compute in the world to satisfy us.
Right.
We asked you to come on, Nikesh, not just because you’re awesome. You’re now a bit based.
He’s a bestie.
Well, he’s a based bestie.
He’s a based bestie.
He was very quiet, under the radar.
I hang out with Nikesh every 48 hours.
I know, David. We keep running into each other.
Every two days. Every time there’s a Knicks-Warriors game, Nikesh invites me to sit with him, so it’s awesome.
Is there one—
That’s not true. I’m just manifesting. Scott Gorin’s seats, I can manifest. A boy from Brooklyn is manifesting. This is who I am.
As long as you’re willing to go with my 11-year-old, you’re welcome anytime.
I’m totally fine with that. I’m sure he—
He’s always looking for somebody to go with, so—
I love it.
—you could be next, Jason.
No, I’m happy to babysit if it’s with the courtside seats or second row.
He’ll babysit you.
Exactly. But there’s a cyber panic going on, all seriousness.
Yes, yes.
2. Cyber Exploits Hit the Web
The last couple of days, AI found two major exploits that impact most of the internet. PHP, which is an open-source coding language, powers about 70% of all websites—think WordPress and Wikipedia—and it’s at risk. In August, security researchers using AI found two critical exploits nicknamed PHP2Shell. PHP’s security team gave this a severity rating of 9.2 out of 10, basically DEFCON 1 for these security teams.
Since August, developers and companies have been working together on fixes. This is standard operating procedure, and the plan was to roll these out gracefully in one major patch on October 22, which is the day we’re recording this very podcast. In the days leading up to the release, information about the exploits leaked. Nobody knows how it leaked. That’s gonna be part of the postmortem of this cyberattack week.
So PHP moved the release up by just a day, to Wednesday, October 21, and on Wednesday all hell broke loose. AWS reported a jump in hacking activity, while Cloudflare’s CEO, a friend of the pod, said they saw millions of hacking attempts as well. A bunch of precautionary shutdowns happened. Government websites across Europe—in France, Germany, the Netherlands, Denmark, Finland, et cetera—started taking websites down. Wikipedia took down some of its systems just to get them patched.
In the United States, the NSA and the FBI put out a bunch of warnings, and now everybody’s started panicking. People are starting to think that Bitcoin’s gonna get hacked, their website’s gonna get hacked, their bank accounts are gonna get hacked. It’s a bit of chaos. And then there’s a second sub-issue around agents accessing banks that dropped at the same time.
Now, you’re in the security business. You’ve got lots of different products and services here. When did you first learn about this, and how severe is it? Is this an acute issue, or is it a harbinger of things to come? And is this gonna happen every 10 days for the next year, as these models have gotten incredibly powerful at what we call cyber—which really means cybersecurity, not just cyberpunk, but cybersecurity?
Oh, Jason, I think 10 days is too long.
Okay.
I think these things are gonna happen much faster than every 10 days. Part of the challenge we’re seeing is that we’ve been talking about technology debt that we all have to go pay sooner than later, and what’s happened is people have not paid attention to the idea that they have to go fix their stuff.
I talked to the CIO of an Azure customer this morning. They had a PHP problem. They have a bunch of websites. Some are running the most advanced piece of software, so they can put the patch. Some are running older versions that they can’t patch. They’ve got to bring them down and go build them up. But they never did anything about the fact that they were running older versions because it wasn’t important. So now they’ve got to take those websites down, upgrade them, put the patch in, and bring them back up.
We found out about this a few days ago. We’ve scanned everything in our infrastructure to make sure there were no issues. We made sure that our products had patches available for our customers who deploy them. There are 2 ways to solve this problem. One, you can go in and patch the vulnerability. You can go in, find the code, patch what has been given to you by PHP, and make it work. The other is you can put it in a perimeter defense product like ours. Cloudflare has one, or other firewalls have one. We put the patch there. It’s a virtual patch, so any attacker trying to access that vulnerability gets stopped at the perimeter because we know what they’re looking for.
They’re coming to a CDN, a content delivery network—
That’s right. That’s right.
—first, and that acts—
Yeah.
—as a first line of defense because they’re not getting to the actual servers.
That’s right.
But if they do get to the actual servers, my understanding is they can run code on these servers, and—
Yes.
—there were some security certificates that were incredibly brittle and—
Yes.
—easy to hack.
The closed models have basically said, “Hey, these things are cyber risks, and so we need to be very thoughtful.” I know Anthropic did that with Mythos, where—
Yes, yes.
—they brought you in and a bunch of other people, and 50 companies, and said, “Okay, now go harden your systems.”
Were the closed models helpful here? Were they the ones that helped you find these CVEs? Bring us behind the curtain.
The closed models can be helpful because OpenAI and Anthropic, and most likely Google as well, have very powerful models now. If you give them enough compute and enough persistence, they’re going to find vulnerabilities. They’re even going to chain vulnerabilities to find an attack path. If they can solve Navier–Stokes, they can chain 4 vulnerabilities, and if you give them enough immense compute, they’ll try every way possible to do it.
Actually, the PHP problem was solved by persistent compute. This problem was identified a few years ago, but nobody thought it could be solved because it required too much coordination. If you throw compute at the problem, it persists and finds a way to crack it. Once it finds a way to crack it and that gets published, or it gets reverse-engineered, every hacker knows how to go after it.
The challenge you have with patches is that hackers are smart. They look at a patch, see what it’s trying to do, and reverse-engineer it from there. They say, “Okay, if I follow the reverse-engineered path and somebody hasn’t patched it, I can break in.”
It seems like a bunch of the hullabaloo today, which seems like it’s going to change every day, is about these bank accounts getting accessed. Do you think that’s primarily through the front end, through a website, using an agent that’s logging in on a browser-based system on behalf of its user? It seems like a lot of this may be coming from Muse or some of these other personal agents that are logging into your bank account to get information about transactions and what you’re doing. The banks are seeing all this activity and saying, “Oh my gosh, is this authorized? Is it not authorized?”
The traditional way banks have provided access to account data has been through APIs using verified applications. I, as a consumer, would log in to my Mint.com account, and it had a legal access point to Amex or Bank of America to—
Plaid is the service there.
It could, yeah.
Plaid is an intermediary.
Yes.
There are a bunch of these intermediaries that provide that capability to apps. Is that ultimately going to lead to a world where the banks all turn off website access and say, “No one can log into a website,” because agents could technically do it and banks aren’t able to manage that?
How do you think this is going to evolve as it relates to consumer financial services? I actually think this is going to be a big part of the press cycle over the next week or two. If your banks are—
Two different things. Yeah, I think those are two different things, David. I think the reason banks have found unauthorized access to their accounts is possibly the PHP problem, because their websites are hackable and people have been able to get through them and get the credentials.
I think there are two different problems. One is unauthorized access happening because these sites have become hackable. The second problem you’re mentioning is a personal agent problem. So far, most personal agents are using your credentials to log in.
Mm.
So it becomes hard for a bank to decide whether it’s actually you logging in or your Muse agent, which has your credentials, logging on your behalf. They can’t distinguish between your traffic and agent traffic. They’re both logging in through the website, basically logging in as you in your browser and trying to conduct the transaction.
As you know, there’s a technology called CAPTCHA, which tries to determine whether it’s a bot or an agent accessing the account instead of you. Today, there are ways to validate CAPTCHA with AI, so you don’t even have to have a human on the other end. So that’s also—
AI’s pretty good at it, yeah.
It’s pretty good at it. Sometimes I have to squint to see what the hell it wants me to type.
Yeah, click all the bicycles in this picture.
Yeah.
It takes you 3 times at our age, with our eyesight. But the interesting thing here is there’s a hybrid as well, Friedberg: your Muse agent, your Grok bot, is running a virtual browser. It’ll pop up the virtual browser and say, “Hey, do this CAPTCHA,” and then take over.
Yes.
So, Nikesh, is this fear, uncertainty, and doubt—the old FUD—by the banks to try to neuter or sandbag these agents from getting to the front door of banking? Because then they could just route to a bank that’s cheaper, better, has better rates, on behalf of—
No.
—of the agent than the customer.
Yeah, I—
Okay.
I think that would be giving them a lot of credit.
Very diplomatic.
I think right now the fundamental challenge they have is a liability problem. It’s kind of like if you call customer support at your favorite bank and say, “I don’t know how $50,000 vanished from my account. I never did this.”
Yes.
Right? And—
What’s this pyramid rhino? I don’t know what this is.
Keep showing up in my dreams 3 times next year.
Okay, keep going, keep going, Nikesh. Don’t be distracted. Don’t be distracted.
The challenge is—
Come on, we have a little fun here. We’re number one for a reason.
The challenge is you’ve given your credentials to an agent. You didn’t even realize it got your credential. It’s living on your browser, living on your laptop. It’s going to get an innumerable amount of credentials. I think the biggest problem we’re going to have is who owns the liability and who has accountability, whether it’s enterprise agents or consumer agents.
So I think you see banks saying, “I do not allow you to interact with me with your agent, period.” They’re yelling that disclaimer from the rooftops, saying, “It’s not going to be my fault.”
Don’t forget, it’s probably not even the cost. The liability is secondary. Banks, we all know this: with credit cards below a certain amount, they don’t want a dispute. They just pay you because the cost of the customer support call is higher than what they’re going to have to deal with. If you start having tons and tons of these agents doing this, and you start calling your bank for explainability, you’re going to create a huge amount of costs. I think that’s where they’re going, to be honest. They need—
You have a unique seat, Nikesh, because you talk to the people who are being hacked, have been hacked, and you help them prevent being hacked with all these suites of services at Palo Alto Networks. I want to take you through a couple of theories people have about AI’s impact on security. You tell me if you agree with them or not.
The first one is: This is going to be largely an issue of who has the most compute. Because the white hats have access to the most compute, they’ll be able to outcompute the black hats—in other words, the hackers. Is that true, or is it not yet true and maybe will be in the future? In other words, Palo Alto Networks or OpenAI, whoever it happens to be, could fire up so much compute to defend systems that a bunch of hackers in North Korea, Russia, China, et cetera, just don’t have the firepower, and they’ll be outgunned. True or not true? When will it be true?
True or not true today? Not true, because the underlying systems aren’t architected or orchestrated in such a way that we can launch these defensive swarms. Everything is in disparate situations. If it were true, then by definition, you could send out a bunch of agents to patch PHP if you wanted.
A company’s got 5,000 websites, and its poor IT guys are going and cataloging each one because they’ve never actually aggregated everything, put it together, and understood their systems. So I think we need to get our infrastructure to a base level of preparedness. That’s where the technical debt lives.
Let me take you to the next thesis people have: persistent computing, reviewing code, and people taking proactive measures more seriously are going to solve this problem. Right now, we’re in an interim period where people have accumulated the technical debt you’ve talked about. AI is going to solve that problem because you’ll have persistent tools from companies like yours and others, or in-house tools, constantly checking systems to find vulnerabilities before hackers do. Is that the end game here? True, not true, true in the short term, true in the long term? Take us through it.
True. Today, almost every large enterprise is busy using MITRE or an equivalent to try to penetrate its own defenses, find holes, and patch them. The challenge, as you’ll appreciate, is that the next model is going to be more powerful. It’s going to find more ways into your infrastructure, so you’re going to be playing this cat-and-mouse game while constantly patching.
I think there’s a slightly different version of this, Jason. For a second, let’s assume you will get breached, because there are so many vulnerabilities out there. You can spend a lot of money patching everything, but the question becomes less about how you got breached and which vulnerability let people through. The question is: How quickly can I find them, and how quickly can I eject them?
Hmm.
Right? So—
Okay. So—
Our advice to people is to build the infrastructure in such a way—which is possible today—that you’re aggregating all the signals and data so you can look for anomalous behavior with that white-hat compute you talked about. Because I have more compute. Go ahead.
Yeah, no. The third thesis is that frontier models are causing a lot of this chaos because they’re gatekeeping the most powerful solutions for cyber defense. The open-source community is giving more functionality to white hats, security teams, and CISOs to solve this, and frontier models need to change how they deploy this technology in some way. Are they causing part of this problem by gatekeeping this technology and not giving the raw power to the white hats?
What did we say earlier about giving too much credit as a diplomatic answer?
Yes.
I think your theory is giving too much credit. Remember, the majority of them are researchers to start with. They’re sitting there and saying, “Wow, this is extremely powerful. What do I do with it?”
Right.
And they’re talking to us and saying, “What do I do with it?” These models that are so far ahead are still trying to figure out how to release them, how to package them, and how much compute they’re going to take. They’re all compute-constrained, so they’re rate-limiting all of us. Even though I have access to the most recent model that’s publicly available, I’m still rate-limited. There are tons of issues going on. I don’t think they’re causing the problem; they’re just trying to grapple with the fact that these things are extremely powerful. What they would like is to figure out how to monetize them.
Hmm.
Because the entire problem—
They want to compete with you?
No, they just need money, bro. They’re not listening to your conversations about compute needs and where the money’s going to come from. They’re not trying to compete. They’re trying to go after profit pools because they want a lot of money that they have to give back to Jensen.
So that’s their main issue. Yeah.
3. Who Owns Agent Liability
Let me ask you all a hypothetical question. I use one of these agentic services. I’m a consumer, and I give it my credentials to my bank account. Let’s use your example: $50,000 goes missing. I don’t know what happened. Maybe I escalate it to the agent I use, you know, Instinct, Muse, Dot. Everybody’s pointing fingers at the other one. Maybe there was a prompt injection. Who knows? Maybe the agent just went rogue and transferred the $50,000. Let’s assume that’s possible. Or $5,000, because it’s under some threshold. Who’s responsible? Okay, Friedberg, who’s responsible? And who should be responsible? Who’s responsible? Who should be responsible?
Well, the bank’s terms of service say you can’t use an agent to access your bank account through the website, right?
Does it say that now?
Yes.
Yeah. Given that those are the terms of service, if the agent hasn’t built in the capacity to recognize that and stop that access, then I believe it’s a software liability issue with the agent provider. Unless it alerts you and says, “This violates the terms of service. Do you want to override my block and do it?” Then the liability shifts to you as a user.
And we have evidence of this, Chamath—not in banking, but in e-commerce. Amazon has blocked these, and Grok will now tell you, “My Amazon shopping bot says, ‘I can no longer check out for you, but I can put it in your shopping cart. Tell me which address and which card you want, and I’ll tee it all up for you, but I won’t hit the buy button.’” It’s being adjudicated in real time right now, and I think that’s probably where we wind up.
Nikesh, what do you think?
You are.
Yeah.
Come on, we all have to take responsibility for empowering agents and giving them our credentials. You can’t say, “I gave my credit card to—”
It’s the only solution. It’s the only scalable solution—
“I gave my credit card to Chamath. He went and burned $20,000 on it.”
Yeah.
Who’s responsible? The bank? No way.
Yeah. Yeah.
Is it Chamath? Yeah.
Yeah.
He is, but I gave it to him.
Yeah. But what if the agent, Chamath, says, “I have been programmed to be maximally helpful”? These sycophantic agents will say, “I found a way. Actually, if I fire up this type of browser and simulate human behavior, I think I can get this done.” That’s what a lot of these agents will do: they’ll try to please you, and they’ll find a way. To me, that would mean the platform providing the agent bears some of the responsibility.
No.
Jason, there’s a word in “agent”: agency. Somebody has to give the agent agency.
Okay.
Who’s giving the agent agency, you or Facebook? I think you’re giving it agency to act on your behalf. If you’re the one giving it agency, you have the liability.
Yeah, I think consumers—
By the way, as a practical matter, it’s the only scalable solution, because what will happen—
I agree with that.
When these lawsuits happen, everybody’s incentive—everybody meaning the corporate incentive—will be to point elsewhere. Bank of America will point to Facebook; Facebook will point to the user. It’s the only way this thing continues to scale like this.
But, Chamath, there is an exception case. I think Jason said it in his words, and you said it too, but we didn’t dwell on it. It’s called the rogue agent.
That’s what I’m talking about. Yeah.
If a product does not behave as intended—
Right.
—then, in my mind, there is platform liability. You said this is what the product does. If you buy any product in the world and it doesn’t do what it’s supposed to do as advertised—if it goes rogue on you—then it’s the person who built the product who’s at fault. And this is where the problem with the brand—
And now we’re in Theranos territory: “Hey, it’s capable of doing 140 tests.” Maybe not really.
4. The Risk of Unfiltered Models
Let me ask you an in-the-weeds question. We incubated a company, Obliteration AI, and what this company is doing—
Nice name.
—and obliterated models, which people might not know about but are going to understand. This company, a little controversially, provides models with no constraints and no guardrails. It’s your responsibility, when you have a red team—a team that’s going out and banging on different vulnerabilities—to build up those guardrails. What are your thoughts on this new, somewhat controversial arms race?
Wait, before you go.
Yeah.
Before you go, I find obliteration unbelievably scary. That’s a vector of risk that’s so underappreciated right now. To your point, Jason, it’s a very long-tail thing. Not enough people understand what it is, but I think this is probably the thing people should understand and think about more than alignment. Sorry, Nikesh, over to you.
Yeah, go ahead, Nikesh. Tell us what you think of these services.
I think this is where we get to the dark side of AI. We’ve had all these debates at the White House. Many of you were present. We talked about responsibility and alignment and making sure we tested them before.
The problem is, you’re dependent on the weakest link. If there’s a company that’s going to give you the raw power of AI without any guardrails, without any controls that are built in or can be built in, then you’re basically handing a nuclear sort of bomb to anybody who wants it, because they can afford it. That’s the problem.
At one level, we’re saying we need global treaties to manage the pacing of AI development for the greater good. On the other hand, you’re saying we can have a company that lets you have this most powerful thing with no control, no guardrails, and no governance. It’s like saying, “Hey, here’s a DIY aspirin kit. Have a good time. You can build your own medication. Have a good time. Sometimes you’ll get it wrong, but don’t worry about it.”
Or just giving somebody some really powerful power tools to build things that make things more secure, like building a wall, and you have to know your customer—
I don’t understand what the business model of your company is. Your company’s basically inducing product liability as a service?
No, it’s providing this to white hats and corporations under very constrained conditions. They share the same logo list as Nikesh and Palo Alto Networks. There are people who want access to these tools in order to do proper testing and make things more robust. You do have a higher bar, like not providing this for free, like Claude was doing, and that was—
But, Jason, are you telling me they underwrite that when you put these models in infrastructure and do testing, they’re not going to go rogue on you?
Well, it’s your responsibility to manage the model, as opposed to giving it to—in this, yeah, you know, if you were the NSA or Morgan Stanley—
But how is that—
—you might need this tool.
How is that different from me going and taking an open-weight model and training it to my needs?
You could do that. This is just a place that has the infrastructure to do it, providing the infrastructure you talked about earlier—who’s going to serve, and that kind of thing. And essentially making it more efficient for you to do this with more constraints, and letting you take the liability.
It’s a very interesting place. Why does technical debt build up? Historically, have we just underinvested? Are a lot of these causes of cyber scares just because corporations deprioritize security and technical debt? Do we just need to change that in this new world, and will that essentially solve the problem if we prioritize safety first?
Nobody wants to work on this stuff. The problem is, when you ship something into production, everybody wants to work on the next greatest thing. Nobody really thinks about maintenance as a sexy idea.
The second thing is, if you actually deploy it inside your own virtual private cloud, the problem is software vendors don’t have one version that’s constantly updated. You have a drift problem where you have a version, Nikesh has a version, Friedberg has a version. They’re all out of sync. Now you, as a software vendor, are maintaining 19 versions because everybody has their own pace and cadence for upgrades and migrations.
Then you have the fact that technologies are constantly changing, and the underlying way things can happen changes so rapidly. But the way you’ve done it represents a sunk cost, and you can’t just rip and replace this stuff. You multiply that by millions of companies and hundreds of trillions of lines of code, and this is why it is like it is.
And, Chamath, also to add to that: the incentive alignment problem. CIOs aren’t incentivized to maintain the most modern infrastructure. They’re incentivized to keep costs low and keep everything available all the time. Every time you want to upgrade something, it introduces the risk of blowing up or going down.
It’s like, “I don’t want to change everything all the time because all my users complain. Every time you upgrade something or change something, it causes friction in the system, and then it costs money.” When was the last time your CIO showed up and said, “Yeah, great, have some more money and a bigger budget because of this new toy”?
Yeah, this is a real tragedy. You’ll see this with, say, the CIA or the FBI. If there are no terrorist attacks, people are like, “We’re spending too much money.” It’s like, “Or we’re spending the right amount of money because we haven’t had another 9/11.”
When they come to us and say, “Hey, we need more because, based on what we’re seeing, there could be another, God forbid, 9/11,” Congress, or whoever has the power of the purse at the time, says, “Yeah, we’re not going to give you more money because we haven’t had one in 20 years.” That’s really the core of this problem: they only get credit—
Well, that’s about to change.
—because we’re going to have a bunch of 9/11s in cyber. Yeah.
Guess what? The phones keep ringing every day, and the conversation is always, “Now, how do I get out of this? I’m tired of getting punched in the face every morning when I wake up.”
People really appreciate the CISO and those groups when these things happen, the systems come back online, and the system’s been hacked. It’s going to be brutal the next year or so. It’s going to be a big transition time. Nikesh?
But I think CIOs could become the next CEOs. Think about it. If you believe that, three years from now, every business function wants to transform itself with AI—let’s go to the positive side, right?
Yeah.
My CFO’s never wanted AI. He’s saying, “Dude, I could do all this financial analysis. Just appoint some people to my team.” My marketing people want them. My HR people want them. Almost every function wants more AI people, which has never happened historically. Everybody wants AI people.
If that’s what’s going to happen, and every function wants the CIO to come and help them with AI or technology, instead of waiting for the technology people, then who’s the person running the company?
What percentage do enterprises spend on security out of their total spend, and what should it ideally be? I’m just curious. We have no insight here.
The insight is that typically, of your IT budget, the cost for cyber is somewhere in the 4% to 10% range. Ten percent is typically where you’ll find financial services and the federal government, because they want to make sure things are secure and robust.
Four percent can be cloud people, because that’s where the hyperscalers take care of infrastructure security for you. You just have to worry about application and software security. So it ends up at 4% to 5%. That’s the range: 4% to 5%, up to 8% to 10%.
Yeah. And that’s going to have to go up. There’s an issue, I think, as you’re pointing out: if there is AI talent in the building, we’ve got to build new products and services for our customers.
Yes.
YOLO.
Well, Jason, there’s a whole new category that we don’t even know how to put our arms around: AI. I was joking the other day, but it’s true. Imagine if I said, “Great, you know what? Go deploy AI in my company, and let’s make sure AI is secure.”
The boss says, “Whatever you do, make sure you can interrogate agents’ chain of thought.” That’s easy to say. That’s hard to do. How do I find all the agents that are running? We haven’t built the AI security harnesses that are needed.
I joke that when they invented planes, the Wright brothers were out there. Nobody said, “Let’s go create the TSA. It’s about time. We’ve got to build security now because bad things are going to happen.” If people are busy enthusiastically deploying AI and embracing AI, you saw—I don’t know if you saw this morning that some researcher found that if you send a Muse conversation to somebody else, you can extract private information out of it.
Yeah.
So now they’ve got to go back to the drawing board and patch it. Everybody’s experimenting, and new things are being found out. We need to build all the security harnesses for AI, and that hasn’t even started yet.
Bernie Sanders goes on Twitter today and says, “Look at all the craziness. This is insane. Stop AI now.” So you’re in a room with Bernie Sanders. Someone says, “Nikesh, we need you to save America. Go meet Bernie Sanders.” You get on the plane, you go over there, you get in the room, and the guy says, “We’ve got to stop AI now.” What’s your response?
How are you going to solve this problem, Nikesh Arora? How are you going to solve it?
I say, “Sounds like a great idea. Let’s go. Why don’t you start? I’m right behind you.”
Yeah.
It’s like, okay. Then, with all respect, it’s interesting. This is the first time in history we’ve dwelled so much on the dark side of technology, and I don’t know why it is.
Maybe you guys can explain to me.
Mm.
We spend so much time obsessing over the dark side of technology. My thesis is that it started with the godfathers, or whatever you want to call them—the builders of this technology. They’ve over-indexed on and dwelled harder on the dark side of it, causing this narrative. I think we need a narrative shift.
Think about it: We only hear about all the bad stuff this technology does. “Oh, my God, I got mid-source. Wait, I got something worse. My guys jumped over. Wait, mine jumped over 6 months ago. Mine are worse.” This is not a good narrative. From a narrative perspective, it’s a race to the bottom.
We’re building a wet lab.
But, Nikesh, why do you think Bernie Sanders and others are seizing on the moment? I know we’ve talked about this before, but what’s motivating Bernie Sanders to be so vocal about “Stop AI, stop AI”? Is it a power struggle? Do they see an emerging system of power they don’t control, that they don’t have a hold on, and that diminishes their power? Or is there something simpler than that?
David, this is where I phone a friend. You guys are much more competent at answering these questions than I am. Jason and Chamath are probably better equipped to answer them than me.
Yeah.
Yeah.
Yeah. It’s a populist issue. If people at the labs are going to tell you half of white-collar jobs are going away, all the drivers are going away, and 10% of everybody is going to die, what a great issue to align yourself with if you’re a populist, whether you’re an America First populist on the right or a liberal on the left. It’s a great way to get people’s attention and get votes. Maybe it helps you win another election.
Nikesh, we appreciate you. You’re busy, and this is the busiest time you could possibly have, I think, right? What is your day like? Have you slept? Are you just getting woken up by clients at 1:00 a.m., saying, “Hey, I need forward-deployed engineers from you”? What’s going on in your life?
Yeah. I woke up at 3:00. I flew to the East Coast. After this, I’m going to get back on the plane and come back home because I’ve got to make it in time for dinner with customers.
Crazy. It’s just a crazy time for you, man.
I love that.
No sleep. I know. Yeah, I love it. Yeah.
I love that.
I mean, this is the time to grind. All right.
I love that.
Thanks so much, Nikesh, for joining us.
Thanks, guys.
We’ve got to get you on for a full episode and talk about a range of issues since you’re so big.
Thanks, Nikesh.
Sounds like a plan.
He does one tweet per day, and it’s based in the cash.
All right. Thank you, Jason.
5. France Faces a Breaking Point
All right. Next up on the docket: Is this the fall of France, gentlemen? We covered this 2 weeks ago. The situation in France is, my God, the place is on fire. It’s gotten worse. Over half a million people have taken part in the protests. They’ve now spread to Italy, Spain, Albania, and a couple of other regions. Over the weekend, France saw its worst riots in decades, and they love to riot, so it’s a high watermark. The numbers are pretty striking. The French are good at it. This is their wheelhouse.
Pro-rioters. Pro-rioters.
They’re professional rioters, exactly. It’s kind of their number-one pastime. For us, it’s football; for them, it’s rioting. Over 3,000 businesses, Chamath, have been damaged. You’ve seen the videos, obviously trending all over the place, with hundreds of millions in losses. Over 1,000 police officers have been injured, and 2,500 people have now been arrested. Seventy percent of them are adult men.
You remember, this was a high school protest because high schools were dilapidated and underfunded, and that’s what the origin of the protest was. Now it’s just become complete civil unrest and looting. Multiple historic government buildings have tragically been burned down, notably the 143-year-old city hall in Limoges, a town 250 miles south of Paris.
Limoges.
Beautiful name, beautiful building. Absolutely heartbreaking to watch it collapse. For what reason? So they could upgrade the schools in France? It seems like there’s a lot more going on here. France has finally implemented some curfews. I don’t know how you don’t do that on day one.
The New York Times reported that France’s government has accused ultra-left militants and opportunistic looters of causing the violence. President Trump and Vice President Vance have called this a third-world immigration and welfare-state issue. Here’s what France’s justice minister said: “Rioters include people from immigrant backgrounds, people who are foreigners, but there are also people who are fully integrated and who are French.”
Polymarket says there’s an 18% chance France declares a state of emergency by the end of the year. At the same time, investors are dumping French debt, Friedberg. France’s 10-year yield has surpassed 5% this week, the highest since 2002. On the other hand, and this is an ongoing issue here in the United States, the 10-year yield dropped just under 5% for the first time in a month. The Wall Street Journal says this might be because people are fleeing EU debt and buying U.S. debt. Polymarket says there’s a 91% chance Moody’s or S&P downgrades France by October 31.
Let’s stop there. Chamath, you spend your summers in Italy, obviously. We talk about that here on the pod. You see it happening in Italy. What’s your takeaway from what we’re seeing?
It’s heartbreaking to see what happens. Let’s just be very clear: I think these people are absolute criminals. They’re all terrorists, and they should all be arrested. I think this shows what’s at the end of all this so-called “empathy” that all these governments have been espousing. It’s just a disguised way of promoting urban chaos.
I don’t know how a police officer thinks it makes any sense for him to be deployed into this melee to effectively protect a Christian Dior store in the 7th arrondissement of Paris. It’s so stupid. These folks have created the tinderbox and the conditions for this to happen, and they’ve done it because they won’t even acknowledge what this is. Even when Macron went on state television, he tried to talk about this in terms of the suffering these people must have. When you look at these pictures, these are not people who are suffering. These are people enacting chaos and violence and terrorism—domestic terrorism. I think you need to respond to these people with an incredibly heavy hand, and then you need to fix the much broader problem.
As you said, Jason, I think France is going to be in an incredibly difficult economic position. We’ve already talked about this before: France is this critical country that’s going to set the tone for the rest of continental Europe in terms of how it finances its long-term debts. It’s already proposed 50 billion euros of austerity. That’s going to have to go up. I don’t understand how, in an austerity scheme, you actually invest in things like education in a way that would appease these people, if that were what they were rioting and protesting about.
The government is in a no-win situation. I think they’re going to lose the upcoming presidential elections, and I think you’re going to see a very complicated reset of the French bureaucracy and the French economy. This is the harbinger of a lot of change to come, because the average French citizen should not tolerate what’s happening in their country right now.
Friedberg, what’s your take here? There might have been completely valid, peaceful protests, with high school students protesting. It’s a beautiful thing. They have a gripe. They created a lot of the democratic system that we actually built America off of. They were great pioneers in democracy and protest. Taking to the streets peacefully is a completely valid form of protest, and then it tips over into something that’s obviously, at a minimum, incredibly criminal and potentially, in Chamath’s interpretation, terroristic. The definition of terrorism is to terrorize people for your political goals. It seems like it fits the definition.
What do you think this represents? Is this a harbinger of more to come, something we might see flare up here in the United States again? Remember, we had the BLM riots, we had the January 6 riots. We’ve had our own series of riots here, and they had a very similar pattern: people who felt politically or socially that maybe they weren’t being represented properly, and then, in some cases, it tips over into criminal looting and incredible violence. What’s your take here, Friedberg?
I think it’s a little conflating to add some of that other stuff to the mix here.
How so?
I think those have very acute political or sociopolitical events tied to them. This one seems to be much more broad-based. If you watch some of the interviews done on the street, there’s a wide range of complaints if you listen to the folks who are rioting.
6. The Welfare State Under Strain
Everything from schools to Free Palestine to the government not doing enough for us with healthcare, support, and jobs. We’re underemployed. There’s just so much discontent, with the objective being to get the government to do more for the people. There’s a lot of immigration that happened, and all these immigrants who come in expect the government to provide for them. The promise was that the government would provide for them.
Over time, the government providing things to the people, as we’ve talked about countless times, is never, ever, ever, ever, ever a sustainable answer. The centralized government doesn’t have accountability. There is no progress, no possibility for economic growth, and no possibility of creating the right incentive structure. So the more the government tries to provide for the people, the more the people are let down.
We’ve seen this in the United States, I would argue, with the 3 things the government is trying to provide here: healthcare, housing, and education. The more the government spends to try to provide more of those things, the worse they get and the more unattainable they become. The same has happened in France. Take a look at this chart. It’s a measure of public spending as a percent of GDP, published by Rand Group Research, and you can see France’s public spending as a percent of GDP compared with peak public spending as a percent of GDP in major communist nations: the Soviet Union during 1989—
Oh, this is an incredible chart, Friedberg. Wow.
—Communist Romania, Venezuela under Chávez, and then Communist China.
That’s unbelievable.
And so, yeah.
So do you think France is a communist state?
Well, this is—
Well, it’s certainly been a socialist state, right?
Is it working?
That’s my point. I think this is such a prescient lesson for the West to look at. We all think there’s some political problem with Macron or whoever it was. The more the government spends, the worse the conditions get for its people, not the opposite. And then the people riot and revolt and say that the government isn’t spending enough and needs to be doing more.
Now take a guess. Pull that chart back up. Take a guess where the U.S. is today.
What percentage of our GDP is—
I’d guess 20%.
Yeah, I was going to say 25%.
Federal government spending is $7.4 trillion. State and local government spending is $4.6 trillion. You add those two together, and it’s $12 trillion out of a $33 trillion GDP. U.S. public spending is roughly 36% of GDP.
Okay. Yeah, we’re right there.
France today is the same as the U.S. in terms of debt to GDP. It’s about 120%. France’s national government spending deficit is 5.4% of GDP. The U.S. is actually at 6.2% of GDP. France is probably the first to go because its economy is growing at half a percent, while the U.S. is growing at 2.3%. So we’re still kind of hanging on because we have some degree of economic growth that’s allowing us to hold things together.
But I would argue that we’re facing the same sort of fiscal cliff that leads to this extraordinary social unrest as France. All the economic measures are aligned: we’re at the same rate, we have the same set of problems, and we’re starting to see the cracks in the system. I think France is a bellwether for the West.
You can load in all the immigration stuff with it, because I do think that’s a big part of it. You open the border to let people in to try to grow the economy, and you promise them things. They actually add fuel to the fire, things get worse, and it becomes a real problem. Fundamentally, socialism destroys progress. It limits everyone’s upside by giving everyone a floor, and limiting the upside is what leads to these sorts of problems, where no one feels like they can progress.
As soon as the quality collapses because you’ve taken out all the incentives in the system, everyone starts to feel incredible discontent. France, I think, is one of the first to go. But if we don’t right the ship in all these other countries in the West, we might find ourselves falling down the same hole, and I’m very concerned about that. I think we need to observe this for what it is and be really thoughtful about it, with our politics becoming, unfortunately, a little more disturbed going into this midterm and over the next 2 years. I worry quite a bit.
What do you specifically think should be done?
Cut spending, cut the deficit to GDP, cut programs. I—
Okay, but hold on.
I’ll give you my 3—
Let me just push back.
Yeah.
Let me just push back for 1 second. What do we think is at the root cause of this? Let’s just focus on the young people. The young people aren’t optimistic about the future.
Yep.
Right? They have a lot of difficulties financially and career-wise, as well as socially. I think their social progression has been stalled relative to other generations. They have these things like swords of Damocles hanging over them, like pensions. I’m sure they don’t understand it completely, but if they took a second to realize that at some point these pension things come due for their parents and grandparents, that’s going to be on them. So how do we set it right for them? What do we do?
I think 4 things are the primary causes of this that could be directly addressed: education, housing, healthcare, and the retirement system. Every one of them probably has a multi-year solution. On education, I think we need to wind down the federal student loan program. I’ve said this many times. And I think in the process, we’ve got to eliminate a lot of the outstanding student debt. I think it’s a necessity.
So cancel the underwriting at the federal level, but also cancel the existing debt of the folks who have it.
There’s a complicated mechanism I think is needed for the deletion of student debt, perhaps in bankruptcy or something like that, to go back on that law that made it impossible—
Just on this one—
—to discharge it in bankruptcy. Yeah.
Just to add to this, I’m extremely sympathetic to canceling student debt. I don’t think it should be—I think it should always have been dischargeable through bankruptcy. But I’m quite supportive of this idea that—
But I think it has to happen with deleting the federal student loan program. What will happen then is you’ll right-size all the institutions and their tuition costs, because all the institutions will start to fit the market. It’ll take about 10 years. It won’t happen overnight.
Mm-hmm.
So we have to have a firm commitment to seeing it happen over the next decade. In housing, we’ve got to do something similar with housing loans that keep inflating the cost of homes and deal with the fact that a good chunk of the middle class—
Okay, wait, but go through your list.
—is going to take a… Yeah.
Go through your list, because I have one critical question at the end of your list. Your list makes sense on the surface.
Then I think we need to make defined-benefit retirement programs, including the Social Security Trust Fund, illegal. I think all defined-benefit programs need to be transitioned to defined-contribution programs over, call it, a decade. In the process, whatever you’re owed in Social Security gets deposited in effectively an IRA or a Trump Account, or whatever they’re called for you today, and all of Social Security is deleted. Now everyone owns equities.
Superannuation funds, like we talked about in Australia. We just transfer to that. What would the pushback on that be, Friedberg?
Because people lose a guaranteed income. But rather than do it all at once, you’d transition 10% of all the Social Security amount owed to you per year for the next 10 years. So you do a 10-year transition where everyone gets their money—
So you smooth out—
—you smooth it out.
—buying into the market. Yeah, great.
And then people can choose a 40/60 plan or an 80% bonds plan, or whatever plan they want, depending on their age—
Do healthcare and housing—
—so they can choose.
Okay.
Healthcare’s the hardest, but I think housing… Healthcare is way more complicated, but I think it fundamentally has to get pushed to corporations that employ people, where they’ve got to go into a market for insurance. That forces the market for insurance to force prices and services down. But there are 2 things that get in the way.
And housing? Wait, and housing?
HIPAA and housing regulations are the 2 things that get in the way of prices coming down in healthcare and housing. So you’ve got to address those, which is much more complicated because, in the case of housing, it’s state and local. In housing—
So let me just ask a question.
Yeah.
When you look at France and Italy and these riots, I think it’s important to observe that they have theoretically 2 of the 4 roughly solved, which is they give you free education and free healthcare.
Yeah.
Okay?
Do you know how cheap it is to go to college in France, Chamath?
I don’t.
It’s 300 euros a year, and it costs the state—
Okay.
—like 12,000. So they’re in college and graduate school. When I was in Paris this summer, somebody explained the resentment to me. They consider it—I’m trying to think of the equivalent we might consider here in terms of entitlements—they believe they should be able to go to college and graduate school forever.
No, but what I’m asking you, Jason—
Yeah.
—and I’ll ask both of you: You bring up these 4 issues.
Yes.
One is a vagary of the United States, meaning the student debt issue is purely an anomalous thing that exists in the U.S. and doesn’t exist in some of those Western countries where we’re seeing riots. The second thing we brought up, healthcare, is roughly free over there. The third thing we brought up, education, is roughly free over there. The fourth thing we brought up is defined benefit versus defined contribution. Again, the system there is a problem because it’s effectively the same system that we have, and you have this pension overhang. But I guess what I’m saying is, 2 of the 4 are roughly solved.
I understand the question.
1 of the 4 doesn’t exist.
Yeah. I can answer.
Yet they’re still rioting. So explain that.
I can answer.
It’s not solved. No, I can actually—
It’s not—
Hold on. I can answer this, I think. The original riots, when it was students, were banal—maybe throwing a couple of bottles, that kind of thing. The students weren’t looting. The looting was carried out by illegal immigrants from North Africa, and there’s a lot of resentment from Islamic Muslim immigrants in France who haven’t been allowed to integrate into society. In some cases, they’re 2nd- and 3rd-generation Parisians who have either not integrated into society, or this is also partly on the native French, who have not integrated into society.
If those people weren’t there, Chamath, we wouldn’t be having this conversation. We’d just be talking about run-of-the-mill people who want a slightly better classroom and a slightly better teacher-to-student ratio. This is because of the unmitigated disaster of immigration to Italy, Spain, and France. People have a hard time just saying it like it is, but that’s what it is. Here, I think we stopped it by closing the southern border. They have a different type of immigration there: people who don’t want to integrate.
If you take the average Mexican who comes to America, they’re more than happy to learn English and integrate into American society and our customs. It’s not the same in the U.K., Italy, and obviously France, as we’ve seen. It’s an immigration issue.
But Chamath, I don’t know if it’s actually solved, because I think the time to get a doctor’s appointment is weeks, and the quality of the education is poor. You’re sitting in a classroom with 60 kids. That’s what the kids are complaining about. There are obviously a lot of other motivations in those riots, but there’s a fundamental feeling of oppression that I think weighs heavily on this entire population.
The oppression is that there’s a higher caste that is doing better than we are, and it’s not fair because we were promised that we would do well. Pull up this video, Nick, that I sent. I think this is a view of some of the anger. “The Chinese have become too rich in France.”
Too rich?
Yeah.
And that’s not fair?
Well, I think it’s not fair. They’re now in the category of people with nice clothes and big cars. This was a response interview about why the rioters were going after Chinese-owned stores and destroying them in Paris. They were specifically targeting Chinese-owned stores, breaking into them and burning them down, because they think these immigrants have become too successful in France and don’t deserve that success.
I think that’s what this weight of socialism, this weight of a giant government, does to everyone: if anyone is seen as getting ahead of the suffering we’re all dealing with, they must be the problem. They’re bad. They’re the issue. I think there’s this fundamental rooting for the Free Palestine movement, on the patriarchy—it’s all rooted in the same sort of principle. They don’t realize that the reason they’re feeling this way is that the government is already trying to do too much and can’t do it well, because the government is the worst place to provide services to the population.
So they see any degree of success, and it’s sort of like what we’re seeing with the populist movement in the U.S.—hating tech, hating billionaires. It’s very common across the West, and it’s going to continue to grow, unfortunately.
Well, there were reports that they were specifically targeting Chinese and Jewish businesses and specifically avoiding Muslim businesses. This was when it moved into the full-scale riots and terrorism—
That’s right.
—and avoiding Muslim businesses.
That’s right.
I’ve got to say, thank God for the Second Amendment. I don’t think we would have these kinds of riots here on a sustained basis. If you remember the L.A. riots, there’s a term called “rooftop Koreans.” The Koreans literally went on top of their businesses and defended them in K-Town in L.A.
When you see these poor business owners unable to defend their businesses, beaten by these mobs, you just think, “My God, that poor small business owner. They put their entire lives into this business.” It’s not easy to run a small business, and then they’re being beaten and watching their life’s work burn down. That’s just unacceptable. I honestly think you’ve got to rethink the Second Amendment as well. I know I’ll get massive blowback, but you don’t see this in Texas or L.A.
Let me tell you how I try to think about what the endgame of this looks like.
Yeah.
Basically, what you guys are saying is that, independent of who’s in charge—it could be left, it could be right—and independent of what the facts are, you showed a chart here that says France spends more than every other communist government in the history of the world. These things are a disaster. Part of the problem is that people probably think this isn’t true, and they want to see even more state intervention to make things better. They know they can’t solve the problem themselves, so the only person to turn to is the government, because it’s not like they’re going to turn to the person on their left or their right.
That’s right.
Okay, so—
That’s right.
—you go to the government. But at some point they’re going to realize that chart is true, and they’re going to see it enough times on television or on TikTok, X, Facebook, and Instagram that they’ll be like, “Wait a minute, I had no idea we’re spending twice as much as communist China. I thought the whole point was that we should be spending more like communist China or communist Russia, because this socialism thing is the better version of communism.”
It’s not going to work. So then who can really fix this? That’s where I go, because independent of what you think is right, who can fix this?
You need Neo in The Matrix.
I go back to 2 things: one is in our control, and one is in the control of the large pools of money and what I euphemistically call these bond vigilantes. The bond vigilantes can break the back of the French government economically by essentially making its debt unfinanceable. The knock-on effect of that would be to push long rates so high that they can’t finance their debt, essentially causing an economic reset inside France.
In that, you have to have a grand bargain where you can try to reset these things. The problem is that every time it’s happened in the past, somebody accrues more power and thinks about their own ability to then have influence over the next 5 or 10 years after that reset. It’s never about resetting in the right way.
The right way would mean looking at the pension system, to your guys’ point; enforcing rules around student-to-teacher ratios in classrooms; enforcing rules around wait times at hospitals; enforcing rules around how quickly or slowly you can get building permits and what you can see in rent-adjusted inflation, and so on. You could do all of these things if you had the political will. Unfortunately, that’s not how it’s going to go down. It’s going to come down to who is in power, who wants to pin the blame on somebody else, and who wants that power for the next 5 or 10 years coming out of a reset. Because in the middle of a reset, what happens is you just give away all your rights.
So that's not the solution. So then what's the solution? Then we're only left with, okay, how do we all solve this? And you're back to the whole question: Is there a panacea? The only panacea that nobody wants to believe is true—we say it because we believe it—is AI abundance and all of this beautiful stuff. It just sounds like meshugas. If I were anybody in France, I'd be like, “AI? What are you talking about? A bunch of oligarchs from Silicon Valley? No, thank you.”
It's almost this intractable problem. It's like the irresistible force meeting the immovable object, and I am completely at a loss to simulate an outcome here in my brain. I don't see a reasonable endgame, and so I'm completely confused and completely without—
Yeah, it depends how far the collapse goes. You could have this sort of social collapse that you see in some countries, involving rioting and murder and all these sorts of things. Or you end up having more of an economic collapse, which is sort of like what Argentina and Brazil, among others, have gone through, and Greece more recently, where the government's inability to service the debt because it got too big forces this kind of restructuring—
Hmm.
—of the economy. And yes, there are still massive protests. There's still an incredible loss of net worth. There's an incredible inflation effect that effectively wipes out asset values for people, and you end up having an economic reset without the great social reset.
So, which path do you choose? I think we have to do something. I don't think we can run a higher deficit-to-GDP ratio than France in the United States. With the same debt-to-GDP ratio as France and only modestly higher GDP growth, I think we're fundamentally going to need some degree of reset in the United States. The question is, do we want it to be a massive social reset, or do the left and the right want to put on their hard hats and say, “Okay, we've got to go in the mines and do the work, and figure out how to cut this from 7.5 a year to 4 a year, and do it in a way that doesn't destroy the economy”? And I think there are a couple of solutions—
Right.
—that will allow that to happen. You can deflate parts of the economy, like the higher education system, by taking away federal student loans. You can have a hit to housing, which will drop the net worth of the middle class by 10% to 20%, by taking the federal home loan system and reducing the underwriting that's happening, reducing the volume that's pushing through. We've used it largely to keep prices propped up and homeowners' net worth high so that the middle class remains highly solvent and feels wealthy.
The problem is it's now inaccessible to young people, so they can't buy a home. We're now hitting a wall here. Something's got to give. If you want social unrest, you don't move on that. But if you want to allow them to feel like they have access, you're going to have to take a hit. The middle class, upper class, wealthy, and so on in the U.S. are going to have to take a hit on asset values and transfer them more carefully to the emerging folks.
Then you've got to fix this pension system, which is either overloaded or underloaded. In both cases, it's really bad. This is why, when ERISA happened, all the private companies in the United States got rid of defined-benefit programs. The only defined-benefit programs in the United States that still exist are inside governments, because unionized government employees said, “We want to keep defined benefits.” If they'd moved to defined-contribution programs, including Social Security, they would all have had so much more money. Everyone would have been wealthier.
You don't even have to wonder about this. You can literally look at what's happened in Australia and how happy people are that they're forced to be in the market.
Norway.
And Norway. They're just like—
And Norway, yeah.
And Norway. If you're forced to do that…
And then we had this incredible thing this week, which dovetails with this: Elon said he's going to give every American child—which I think is defined as under 18—a share of SpaceX worth $150, $160, $170, which is what it's been trading at. That's a $10 billion gift, which I believe will be among the largest gifts in the history of charity. And he said he's just getting started with that.
There's nothing that says this isn't going to become a bit of a contagion among the affluent, with people giving a large portion of their wealth, or even a small portion of their overall wealth. But giving it directly to kids who can't touch it for 18 years—and if we made it part of their superannuation fund, maybe we could have Invest America up to age 18 for college, and then we could have an Invest America for your retirement. If you bifurcated that and gave one share to each, that could be even more accretive, right?
Jason, what did you think about your friend Kara Swisher's criticism—
No longer friends, Karen. Kara and I haven't been friends for a long time. I think she is one of our—
Can you just describe for the audience what her criticism of Elon was and your reaction to it?
I think she had multiple videos. One of them is that it's only one share, and that they should just give money directly to people. The reason that's incredibly stupid is you should give any possible gift that people are willing to give, and forcing people to hold onto it makes them part of capitalism. Kara Swisher makes $10 million a year from her podcast. She talks about it all the time. She is in the 0.1% or 0.01%.
She should give half of her podcasting revenue to Invest America accounts if she's so virtuous. And she's also saying they should be putting Elon in jail, and that the Democrats are going to do that if they win. So there's kind of a double thing there. But she's a Karen, and she's a hypocrite.
Yeah.
Can I say something about this? Nick, can you pull up that image of that woman whose store was destroyed in France, who's a well-known activist?
Yeah, this is—
—you guys see this image and this direction—
—I did see this, and I don't know whether it's true or not—
So I think—
—but the story here is—
Yeah, the story is that this woman is a well-known activist in France who, for a long time, was very vocal about opening the borders, letting the immigrants come in, and taking care of them. She was angry and upset, yelling at people in the street while they were destroying her store, saying, “Do you know who I am? I am an activist who advocated for you all.” And they didn't give a shit. They went in and destroyed her store.
The look on her face, like, “What the hell happened?” is meant to be this moment that captures the zeitgeist of what some people call suicidal empathy. But I think it's much more about being an armchair do-gooder. An armchair do-gooder is someone who sits at home, in a nice home in the suburbs or wherever they live, and says, “Let all the people in. Take care of all the people.” But as soon as the people show up at their doorstep and say, “Can I come and stay in your house?” the answer is, “No, of course not.”
Yeah.
Someone else has to deal with it. Someone else has to hold you in their home. Someone else has to house you, feed you, and give you a job, because I can't house you, feed you, or give you a job. I think this has become inherent throughout the West: the armchair do-gooder says we should do all these good things, but when the cost of doing those good things comes home to roost—
They're out.
—they're shocked.
Yeah, they're out.
And they're awed, and they're like, “Holy shit.” This, by the way, we see in the U.S.
We had a term for it, by the way.
There are all these great videos on Twitter where people go around to suburban homes and ask, “Hey, did you vote for a bill to have unrestricted immigration?” They say, “Yeah.” They're like, “Okay, here are 15 guys. They want to come live in your house.” And the people are like, “No, they can't live in my house. What do you mean?”
Yeah.
I mean—
This was called, in another era, “limousine liberals.” It was manifested when they used immigrants—which was quite offensive, but quite effective—and bused them to Martha's Vineyard or other places and said, “Here's your immigrants. You said you wanted them.” Yeah.
What Rob Henderson called “luxury beliefs.”
Luxury beliefs, limousine liberals. If you were to—
Yeah.
—look at the problems in France, they're not completely intractable. A lot of the complaints are, I think, wait times at a hospital. Well, you have free healthcare. Oh, my, the student ratio. Well, you're only paying 300 euros a year for your college.
If you want to go from 60 to 1 down to 30 to 1, okay, yeah, we’re going to need to charge you 2,000 euros a year.
No, you don’t. No, you don’t.
Yeah.
You don’t need to charge them more. This is the point. When the bureaucracy inflates and the quality of the service goes down, an intelligent person would look at that inflated bureaucracy and say, “This is inversely correlated to what I expected.” So really, people should have the intelligence to say, “Hold on a second. I’m now spending 57 cents of every gross productive dollar on a government that has increased my wait times and decreased the student-teacher ratio?”
Oh, yeah, they’re certainly not going to be—
I want the opposite.
—not as efficient as the free market.
So—
So, exactly right. That’s right.
But if you have an immigration rate where you’re importing 1.5% of your population a year and your GDP isn’t growing, that might be a signal that maybe you have to change gears. And if the people you’re immigrating into the country are not productive, or not productive yet—maybe their kids will be productive, or their kids’ kids will be productive—but right now they’re a drain on society, you can only accept so many people.
There’s another nuance to this: You cannot accept people who do not make a commitment to assimilate into that country’s culture, period, full stop. You can have a very small number of people, whether it’s Japan, the United States, or France. You can only have so many people join the country who don’t speak the native language, because that becomes an incredible burden. Now you have to create an entire second education system and healthcare system where people speak Spanish, French, German, or whatever it happens to be.
That’s what’s taxing these countries. I think 1 in 19 are from North Africa, and 1 in 9 is an immigrant. They have unmitigated immigration in Europe. That’s the other piece of this problem. If they hadn’t had that for the past 20 years, and they weren’t adding so many low-value immigrants as opposed to high-value immigrants—high-value being people with graduate degrees who are investing in companies, versus people who are taking bottom-rung jobs—and if they started looking at it against the GDP of the country, against unemployment, and asking where an immigrant fits in terms of the jobs we need. Do we need this type of worker, a construction worker? Yes? Okay, we should add more of those because we have a deficit. But they’re not doing it intelligently, and that speaks to this woman running her small shop. She wants to allow all these people to come there, but she’s not looking at the balance sheet, and the balance sheet doesn’t support it.
I’ll simplify it down.
It’s an economic issue.
Yeah. Just to restate what you said slightly differently: If you’re providing social services to an immigrant population, then that’s not necessarily a productive immigrant population in terms of being able to contribute to growing the economy and supporting civil society.
Yes.
Social services that are requisite for an immigrant population should indicate to you that that’s probably not the immigrant population you want to bring in. This has nothing to do with the countries or the character of the people—
Ethnicity, color of skin—
—ethnicity. It’s simply a statement that if you’re paying for social services to support an immigrant population, something’s wrong, and it’s the wrong population that you want. But if immigrants come in and they’re productive, meaning they can get jobs, work, and support themselves, that means they’re adding something positive to the economy and civil society. That’s the sort of selection criteria that needs to take place. So that’s a simple framework for determining whether or not you should do this, and an easy way to solve it is to get rid of social services for immigrants.
Yes.
As soon as you get rid of social services for immigrants, then you self-select into the right population that’s going to come in.
The second thing is, Jason, just to put a point on this: It’s not about assimilating to society. It’s about sharing the values of what makes the United States the United States.
Well, that’s what I mean.
The value of being self-sufficient. I don’t care about your religious beliefs. I don’t care about your movie interests. I don’t care about your first language. I don’t care about your sports interests. The multiculturalism of the United States actually adds an incredible amount of value, but the values we care about—the principles we all share as a civil society of the United States, or a civil society of France—are screening criteria that seem to have been thrown out the window as we’ve allowed the immigrant flows to kind of take over the West.
Yes. Do you believe in freedom of speech? Do you believe in the Second Amendment? Do you believe in the core benefits of America, that you pull yourself up by your bootstraps?
Do you believe in women’s rights? Do you believe that women are equal to men? As a simple one, right?
Yeah.
Do you believe that you should work for your money as opposed to take it from the government? Do you believe in your individual—
Do you believe in the legal system versus your church’s legal system?
Is your church paramount to the government and the laws of the state? And do you believe that other people have rights that you don’t get to trample on?
Correct.
Those are the sorts of things that, in many cases, are counter to the cultural, religious, and other traditional beliefs of certain populations that end up causing the damage in the West. I think that’s been completely lost in the process.
You know what I thought when I heard that kid talking about targeting these Chinese people? It’s weirdly reminiscent of a more bottom-up, maybe poorly organized version of Kristallnacht, where—
Yes.
—you had the looting of Jewish businesses and synagogues in 1938, which were state-sponsored. They had some idiotic pretext, but ultimately it was because they were jealous of the Jews. And now you see folks essentially burning and pillaging businesses, and at least some narrow percentage of them were doing it to target a cohort of people they’re jealous of—and they’re jealous of exactly what you said. I don’t think the Chinese emigrate to a country and have some of these long-tail views. They’re like, “Get me in the game, coach. Tap me in. I’m ready to go.”
They’re entrepreneurs.
And they’re entrepreneurs. They’re industrious. They’re super hardworking.
Exactly.
So the contrast is quite stark. You see one immigrant essentially saying, “C’est pas juste,” which is basically his way of saying, “It’s not fair,” which is kind of like a petulant child.
Right.
On the one hand, you have petulant children. On the other hand, you have hardworking immigrants. It’s the craziest dichotomy. Just seeing it in real life is shocking.
Immigrant-on-immigrant violence. One is working harder, has a better plan, is applying themselves, and is probably working 7 days a week, 12 hours a day at that business to make it successful.
Well, no, her specific story is even more incredible because—
I was talking about the Chinese woman.
Yeah, the Chinese woman.
Yeah. I grew up in the restaurant business, and let me tell you, it’s the hardest thing on a family to run a business, because when somebody doesn’t show up for work or you don’t have the money, you all just have to work. My family, we all worked at the restaurant. My mom made the desserts at our restaurant because she wanted to save money, and she was a nurse, not a pastry chef.
My brothers and I worked at the restaurant since we were 6 or 7 years old. We set the tables. We polished the silverware. We were working in the kitchen at 10 years old, and we were still living week to week. And then you come burn the business down? You come burn my dad’s business down, we’re going to unload. My dad was packing, you know? He was strapped. You do this in America, you do this in Texas, you’re going to get shot.
When I saw those poor Chinese people on the balcony, their lives at risk and their whole livelihood at risk, I was just like, man, I wish these people had guns.
I wish these people had guns and could defend their private property. It is crazy that they could just be run over like this. They were going to get killed. There’s absolutely a nonzero chance that when mobs get into that heightened state, they will murder people. They will beat them to death. This poor other shop owner was getting kicked and beaten just trying to defend his little tiny business, his little tiny store.
If you’ve ever been to Paris, you know those little tiny stores—beautiful, bespoke stores trying to make a living in something they’re passionate about. It’s absolutely heartbreaking, and it’s not—
The bond—I’ll just say this. The bond vigilantes need to show up here and draw a hard line in the sand. I think they need to—
Love it.
—break the back of the French government.
Okay. There it is. There’s the mandate.
And you need to do it severely, and you need to do it promptly—
6%, 7% would do what? 6%, 7% on these bonds?
I think you need to take France to 8% and just break them.
Yeah, and then all of a sudden, austerity measures, immigration, all this stuff’s going to be taken a lot more seriously.
At least it’ll all be on the table, and they’ll have to sort out what they’re willing to give up. But right now, if you can skate through this moment, these things continue to fester, and it is a tinderbox that at some point—again, we know, just to use the analogy of Kristallnacht, we know what the trail of breadcrumbs can lead to in terms of fomenting societal rage. You cannot allow that to happen.
The government has a responsibility now to nip this in the bud in a structural way. And I think the people have a responsibility to understand the numeric reality in which they are living. They are living in a state that is spending more than any communist utopia they hoped to achieve, yet the outcomes are broken, which means the system is broken, and you need to reset that. The government will not reset it on its own, so the bond vigilantes will have to do it for them. This kind of stuff, if it continues to compound, is extremely, extremely dangerous.
Yeah. Super dangerous. We had our little flashes here, so we’re definitely not immune to it. We’re a larger group of 50 states. And, by the way, I think France is almost at 56% as its top tax rate. Maybe it’s 55%, 56%. The only place that’s higher in Europe, I think, is Denmark, which has about a 60% top tax rate.
It doesn’t matter. Take it to 90%. Nothing will get better, folks. Do whatever you want.
If the government provides the service—
Nothing—
—there’s going to be no innovation—
—will—
—and there’s going to be no—
—change.
—and there’s going to be no efficiency. It is what it is. Yeah.
Unless there is an individual human incentive to do well, nothing gets better.
All right. Capitalism at its core. The thing is, Macron was trying to get capitalism going there as well. It was a big part of—
I think you can just—
—his—
—click on this image here or this link, if you want to see—
What do we got here?
This is sort of another indicator of why GDP growth is laggard, but this is tax revenue as a percent of GDP. France also tops the list on tax revenue as a percent of GDP at 46%. So every dollar of GDP generated by the country is collected as government tax revenue. The U.S. is at 22%.
Obviously, 2 drivers here: high tax rates or low GDP. And we know—
No, it’s just the tax, because the tax rate is always a function of income. So—
I think the point is—
Yeah.
—if you put this and the other one together, you have a confiscatory government that represents the overwhelming amount of the economy, where you have incredibly deleterious outcomes. That is the connect-the-dots, and you can’t hide.
This is the question for the United States: Which direction do you want to go? Do you want to go up the chart or down the chart on both charts that we looked at? Which way do you want to go? The quality of life goes down as you move up the chart for all people, and the ability to progress gets destroyed as you move up the chart.
As you move down the chart, each individual’s ability to progress in life improves. You have to do it on your own; you have to have more agency. But your capacity to improve your life improves. As tax revenue goes down as a percent of GDP, and as government spending as a percent of GDP goes down, your opportunity as an individual—the basis of the foundation of this great country—improves. That is really the North Star we should all be talking about: How do we get back there? How do we continue to move back in that direction on the chart?
Pull it up again.
There are very practical things. There are very simple things to do, like we talked about on the show.
Hold on. Pull the chart up again, and I’ll give you a little bit of a challenge here. I think most people in America, especially young people, don’t believe what you’re saying. If you look at the U.S., ahead of the U.S. are Australia, Argentina, South Korea, Japan, Brazil, Canada, the U.K., Germany, Italy, and France. Most people want to live in those countries above us. Most Americans who are complaining now certainly want to live in one of those countries, have healthcare, and have cheaper, more affordable college. Housing is still screwed.
Not most. No, I disagree. Not most.
Oh, you disagree that most Americans—
Yeah, I disagree.
—don’t want universal healthcare?
No, no, no. Most Americans don’t want to live in a country above us on this list. Absolutely not.
Well, they certainly don’t want to go below us to South Africa, Russia, Turkey, China, Mexico, India, and Indonesia.
Well, I’ve been to—
All due respect—
—I’ve been to every single country on this list above us. I’ve spent a lot of time there. I’ve been a citizen of one of them.
Mm-hmm.
And I would not choose to be in any country other than the United States.
You’re an entrepreneur. Yeah, I mean—
Jason, do you know what youth unemployment is in Brazil?
Brazil, I do not know. I’m going to guess 20%.
Youth unemployment in Brazil is around 15%, 16%.
Yeah, that’s—
Of almost all—
—that’s right in the streets category.
—and it’s rising.
Yeah.
And it’s rising. It’s almost at 20%. The U.S. is still fortunately below 10%.
8%. Yeah, 8%, 9%.
Or, yeah. The same is true in many of these other countries. There just isn’t a growing economy that can support a youth employment system in many of these countries. So I don’t know if I would agree with your statement that most people want to go live in those other countries. Because if they wanted to, they could use a credit card, buy a plane ticket, and go do that.
Well, let me just say—
I think the United—
—most people want what those countries deliver. Most people want what—
They want what those countries say they deliver.
—and they would be willing to pay more tax for it.
But it’s not—
Or they’d certainly be willing for us to pay more tax to achieve it.
—it’s not in full scope. The full scope of what those countries deliver is not just an acute, no-fee trip to the doctor’s office. People want to have the ability to do that, and also have a job, and also have the ability to progress. To have all of these things, fundamentally, the best system is a smaller-government system, because the more the government grows, the less of that pie of things you will actually get. It gets worse and worse.
I’m in line with your argument. I’m just saying that chart is not a good piece of evidence in this argument. I think most people are not entrepreneurial, and most people want—
Not about being entrepreneurial.
No, but I mean—
Yeah.
I think more freedoms and less tax are what we all want. I think a large percentage of America wants us to pay more taxes and get more services, and they believe those countries right above us—
Never going to happen.
—want more—
Never going to happen.
—will give them more services.
Never going to happen.
Therefore, they don’t mind waiting an extra couple of hours or months to get healthcare, if they get healthcare, because they currently get no healthcare, or they get bankrupted by healthcare. That’s how dysfunctional our healthcare system is in the United States. I think most Americans would disagree with us looking at that chart. That’s just my belief. But we’re here to sharpen each other; iron sharpens iron. I think we have to keep working on this value proposition.
7. The Minimum Wage Fight
I don’t know if you guys saw my tweet this week about how President Trump could save his second term. A big part of it, I think, is he has to talk about housing as much as he’s talking about data centers. Raising the minimum wage is a nice gesture, since it only impacts 10 to 15 million people at the federal minimum wage in this country.
And there are things we can do to make it apparent to the people in the bottom half of the country that we care about them, and that the President is signing executive orders for them as well as for us. And by us, I mean the tech industry and large businesses.
I think that’s the big problem with why the President is polling so low—his ratings and approval are so low. Perhaps we’ll see what happens with this election on November 3rd. A big part of the discontent right now is that people are frustrated that President Trump doesn’t even talk about them anymore. He doesn’t talk about minimum wage, housing, or inflation. He just says, “It’s all going to be fine,” and that we’re living in the golden age. But then he’s more than willing to talk about our issues: data centers, crypto, et cetera. The messaging in this country is really far off.
Until we get somebody in that office who is willing to talk about the bottom half of the country and strike a 50/50 balance—“Hey, we’re going to talk about big tech, we’re going to talk about lowering taxes, but we’re also going to talk about raising the minimum wage. We’re going to talk about a plan to build more affordable housing, as opposed to just more data centers”—that’s the disconnect—
Well—
—that I’m just bringing up here.
Hold on. Hold on. If you actually read what he tweets, I think he connects the dots on all of these things. I think he has acted in a way that one should when you realize that the AI build-out is a disproportionate percentage of all GDP. It’s basically all of it. And it doesn’t just pull through stock performance for seven companies. There’s been a bunch of data showing that it pulls through massively increased wages for middle-class and skilled labor.
Then he’s been telling people that if they’re going to build these data centers, they should do it in a way that lowers the electricity cost: build fully behind the meter and invest in the local communities. That also has a positive impact. He’s trying to make sure he lands the plane on the U.S. economy. And right now—
I agree he’s doing that.
—whether you agree with it or not, the U.S. economy is entirely and completely levered to and correlated with the AI trade. Period.
Yeah.
I agree that he is doing a great job—in that, he’s better than any president I’ve ever seen in my lifetime in terms of getting the economy rocking.
The overwhelming number of Americans now do not agree with what you just said. They believe that he does not care about the issues they care about. That’s why his approval rating is at a historic low. So there’s a messaging problem here. I’m not saying I should be the press secretary here—
Well—
—but somebody’s—
Well—
—got to be a press secretary here—
Now—
—based on communications—
Well, look, first of all, let me say something about the press secretary, just for one second—a small deviation. We have not spent enough time acknowledging what more people should acknowledge more often, which is that Karoline Leavitt is an absolute GOAT. She set the bar so high. She’s so talented. For being so young, it’s unbelievably impressive. It’s almost an impossible job to follow her. Let me just say that, because I don’t think people say that enough.
Second, I’ll go out on a limb and say this: I think you are an incredibly dogged bulldog. If you got to know the President, you’d be besties with him. I really think you would, and I think you’d be an incredible press secretary. Now, the problem is—
I would be an asset.
—the problem is that you have stupid ideas. And as long as you could avoid representing your stupid ideas—
Raising the minimum wage is a stupid idea? I don’t think so.
No, no, no. You are—
It is a bad—It’s a terrible idea.
No, it is not.
I agree with that. Yeah.
The federal minimum wage is not.
Yeah. Look at the studies.
You are the epitome of an armchair do-gooder, okay? If you could quell that instinct and actually be his vessel, you’d be incredibly effective because you’re dogged, you’re great on your feet, and you’re super funny. You’d be incredible.
People are pitching me to do it now.
I don’t think you’ll be able to manifest this one, but you are incredible.
We’ve been number one the last 2 weeks in a row, so I wouldn’t underestimate my manifesting powers. I do think, Freeberg, if you look at the studies, you’re dealing with old data and maybe a little bit of dogged libertarian principles.
Dogged libertarian principles?
Listen, I share some of those libertarian views where I’m like, “The free market will work it out.” Only 15 million people in this country make under $15 an hour and are impacted by the federal minimum wage, which hasn’t moved in decades. If the President came out and said, “Hey, we’re going to increase the federal minimum wage by $1. It basically doesn’t count for half the states, because they have their own minimum wage. We’ll add a dollar a year for 7 years and get it to $15,” it would just be a good-faith thing to do. All that money gets recirculated and creates more monetary velocity in the country. It all gets spent in the economy. It’s a tragedy of the commons: one person doesn’t want to pay this, and another person doesn’t either.
Disagree.
Look at what happened in Australia. Look at the Nordics. When they raised—
Jason.
Yeah.
California passed Assembly Bill 1228 in 2024 to increase the minimum wage in California.
For fast-food workers, to $20. It was too—
Twenty bucks.
—it was too fast. I am not advocating going from $7 to $20.
So when they did that—
Yeah.
—we lost 18,000 fast-food jobs in the state. The reason is—
And we still have—
Just think about this.
—record employment.
The reason is that the fast-food industry sells food to consumers. Consumers want cheap fast food.
When you increase the labor rate against the market’s demand, the consumer says, “I’m not paying $8 for a burger. I can only afford $4.” So there’s actually no one to pay the wages of those fast-food workers. Number one, the consumers lose because they’ve lost their local fast-food joint; they can’t afford the new fast food. Number two, the employees who were voluntarily and willingly working there for $18 or $16 or $12, whatever the number is, suddenly lost their job.
Yeah, so let me—
And as many ways as you want to slice this—
Let me address it.
—this is not a dogged libertarian principle. This is just a very common-sense view of the world.
The buyers are saying, “Hey, I’m willing to pay more,” and the workers are saying, “I need to get paid more to work here.” The market finds a happy medium. But when the government intervenes, it blows up the market. It destroys the supply, it destroys the demand, and people lose jobs. That’s what we see time and again when the government intervenes in market dynamics.
Here’s where you’re wrong. Number one, I never said to go to $20 an hour. I said to add $1 per year for the next 7 years, which will get us to $15 in 7 years, putting us in 2032. You’re comparing California’s most radical program, which I don’t agree with, to my common-sense one.
Now, on the second point, it has added essentially 25 cents to each meal at those places. It makes them get rid of jobs that are going to be replaced anyway, like cashiers. But we still have the same unemployment rate in these places. People went to other productive jobs when those jobs became automated. That’s an argument you and I have made here many times: people go find work. So you can’t have it both ways, Mr. Friedberg. You can’t say—
I still disagree, Jason.
—that people will find more productive work. The people who lost their jobs found better work, statistically, in California.
You’re trying to be a market prognosticator, determining what the right price for the market is. The more you do that, the more you’re going to destroy the market. Fundamentally, that $1—or whatever you do to rationalize the government’s intervention—is going to have some effect on some part of the market that tips it past the point where it no longer makes economic sense for buyers to pay the premium price.
And some number—
Yeah.
Of those sellers will get—
Out of the market, and the thing goes away.
This would be true if you applied this in an unconstrained fashion. In a very constrained fashion, raising the bottom practically just means those people go spend the money with the same companies they’re working for. The Amazon delivery person or the UPS delivery person already has a $20-an-hour job. But the people who are doing those jobs just spend it at the same companies.
So in reality, when you force everybody at the same time to hit a minimum labor dollar amount—and we already have it, by the way, for this very reason—it just gets recirculated. It’s actually good for companies, and we saw it in Australia, New Zealand, and the Nordics. As they raised it, all the companies did better because their customers had more money to spend, and people were happy.
In a society, you don’t just want to do things for economic reasons; you also want to do it to make people have a better life. I think that’s what raising the minimum wage does. It creates an esprit de corps. It makes everybody feel they’re in it together, and the President of the United States is looking at that lowest-paid worker and saying, “You know what? We’re going to ask Amazon, we’re going to ask McDonald’s, whoever it is, to just do a little bit better for about 10 million Americans.” It has no effect. It’s only 10 million Americans. That’s why I think it’s a no-brainer. It’s a no-brainer.
We’re not going to agree on this.
Yeah, no problem.
I think an alternative would be to drop taxes for the lower bracket and give people a zero tax rate.
Okay.
Or those sorts of mechanisms.
Zero tax rate up to $150K. That would solve everything.
And when you do that, you don’t have to raise the minimum wage.
Agree.
You can give those people that money in their pocket—
They save 25%.
—because they don’t have to pay money out of their pocket. And you can do other things, like credits for them. So there are a lot of ways, Jason, to get that—
Right.
—without forcing your way into the market, which I think always has adverse consequences. You can do that through tax policy and incentives and other structures. But—
Well, we already have—
—let’s agree to disagree and move on.
We already have a minimum wage, so what we’re debating here is what it should be.
Love you boys.
Rain Man, David Sacks.
And instead, we open-source it to the fans, and they’ve just gone crazy with it. Love you, Wes.
I think the queen of quinoa.
Let your winners ride. Let your winners ride. Besties are back. Let your winners ride. That is my dog taking a piss in your driveway, Sacks.
Oh, man. Oh, man.
My Abbe Dasha will meet me at Glenside.
We should all just get a room and have one big, huge orgy, because there’s just this sexual tension that they need to release somehow.
Wet your feet. Wet your feet. Wet your feet. Wet your feet. We need to get merch. Besties are back. I’m going all in. I’m going all in.