Debt Spiral or NEW Golden Age? Super Bowl Insider Trading, Booming Token Budgets, Ferrari’s New EV

Summary

The All-In Podcast core four reunite to cover a packed agenda spanning AI adoption, prediction markets, the national debt, the economy, and Ferrari’s new electric car. The episode opens with a deep discussion on AI acceleration, triggered by a Harvard Business Review study finding that AI tools intensify work rather than reduce it. Sacks argues this validates his prediction that AI will increase demand for knowledge workers, while Jason shares that his firm now has AI “replicants” handling 20% of investment team work. Chamath raises a provocative question about whether on-prem computing is making a comeback, arguing that enterprises cannot afford to leak proprietary data to cloud-based LLMs, especially after a judge ruled there is no attorney-client privilege for content processed through cloud AI tools.

The discussion of prediction markets centers on the Super Bowl, where over $2 billion was wagered across platforms. The hosts debate whether insider trading on prediction markets can or should be policed, with Chamath drawing a compelling parallel to Warren Buffett’s pre-and-post Regulation FD returns, showing that Buffett generated zero alpha once information asymmetry was eliminated. Friedberg delivers a grim analysis of the new CBO report projecting $56 trillion in debt by 2036, warning that the real crisis will come when state and local pension obligations get federalized. Sacks pushes back, arguing the CBO assumes unrealistically low GDP growth of 2.2% when the economy grew over 5% in Q4 2025, and suggests we may be at the beginning of a new golden age comparable to the late 1990s. The episode closes with a fun segment on Ferrari’s first all-electric vehicle, designed with input from Jony Ive’s team, which has gone viral for its stunning interior design.

Highlights

”When do tokens outpace the salary of the employee?”

yt-dlp --download-sections "*16:40-17:40" "https://www.youtube.com/watch?v=CnaegIpkenA" --force-keyframes-at-cuts --merge-output-format mp4 -o "CnaegIpkenA-16m40s.mp4"

“We hit $300 a day per agent using the Claude API like instantly and that was doing maybe 10 or 20%. That’s $100,000 a year per agent. When do tokens outpace the salary of the employee? Because you’re about to hit it. I’m about to hit it. I think superstar developers are already there.” — Chamath Palihapitiya, 16:40

”Is on-prem the new cloud?”

yt-dlp --download-sections "*10:20-11:20" "https://www.youtube.com/watch?v=CnaegIpkenA" --force-keyframes-at-cuts --merge-output-format mp4 -o "CnaegIpkenA-10m20s.mp4"

“Is on-prem the new cloud? Which is weird to think that could even be possible. Once you use these tools, it is very difficult for a company to control how their data is used. If you’re using a set of agents to act on all that information, all those agent traces are going back to these model builders.” — Chamath Palihapitiya, 10:20

”Warren Buffett generated zero alpha after Reg FD”

yt-dlp --download-sections "*23:40-24:40" "https://www.youtube.com/watch?v=CnaegIpkenA" --force-keyframes-at-cuts --merge-output-format mp4 -o "CnaegIpkenA-23m40s.mp4"

“Warren Buffett’s returns were double the market returns when this kind of information sharing was legal. And the minute that it became illegal and you had to act on the same edge as everybody else, his returns went to the market return. He generated zero alpha.” — Chamath Palihapitiya, 23:40

”The concrete that breaks the camel’s back”

yt-dlp --download-sections "*35:50-36:50" "https://www.youtube.com/watch?v=CnaegIpkenA" --force-keyframes-at-cuts --merge-output-format mp4 -o "CnaegIpkenA-35m50s.mp4"

“If you end up with a Democrat-controlled House and a Democrat president in 2028, you’ll very likely see a federalization of state and local pension obligations. When you add that liability to this CBO report, I think that could be not just the straw that breaks the camel’s back, but the concrete that breaks the camel’s back.” — David Friedberg, 35:50

”The beginning of a new golden age”

yt-dlp --download-sections "*55:40-56:40" "https://www.youtube.com/watch?v=CnaegIpkenA" --force-keyframes-at-cuts --merge-output-format mp4 -o "CnaegIpkenA-55m40s.mp4"

“When we look back on this period, it could end up being a little bit like the late 90s. We had phenomenal economic growth, the internet. I suspect we’ll look back on this time period as the beginning of a new golden age.” — David Sacks, 55:40

”The Ferrari experience with Jony Ive’s touch”

yt-dlp --download-sections "*63:50-64:50" "https://www.youtube.com/watch?v=CnaegIpkenA" --force-keyframes-at-cuts --merge-output-format mp4 -o "CnaegIpkenA-63m50s.mp4"

“The screen looks very Mac-inspired except unlike Tesla which is no buttons and removing buttons, they’re adding buttons here and making the buttons very tactile. All the sports car enthusiasts love tactile memory-based buttons that you can just have fun with and flip and feel like you’re a fighter pilot.” — Jason Calacanis, 1:03:50

Key Points

• AI Intensifies Work (0:23) - HBR study by UC Berkeley researchers found that AI tools at a 200-person tech company made employees work faster, take broader tasks, and work more hours, but also increased stress and burnout
• AI Increases Knowledge Worker Demand (1:08) - Sacks says this validates his most contrarian 2026 prediction that AI would increase demand for knowledge workers, not eliminate them; workers get upleveled from task-based to purpose-based jobs
• Enterprise AI from Bottom Up (2:30) - Sacks predicts massive enterprise AI adoption this year driven bottom-up by early adopter employees bringing consumerized tools, not top-down transformation initiatives
• OpenClaw Revolution (4:30) - Jason reports 5-10% of his firm’s work moving to AI “replicants” weekly; they have 4 AI agents with their own Notion, Slack, and Google accounts plus a meta-agent called “Ultron” managing the others
• On-Prem is Back (10:20) - Chamath argues enterprises will shift back to on-prem computing because AI tools leak proprietary data to model builders; a judge confirmed no attorney-client privilege for cloud-processed content
• Token Budgets Exceed Salaries (16:40) - Chamath reveals hitting $300/day per agent ($100K/year), approaching the point where token costs outpace employee salaries; superstar developers may already be there
• Super Bowl Prediction Markets (19:19) - Over $2 billion wagered on Super Bowl; anonymous accounts correctly predicted halftime show details; Israeli soldiers allegedly used classified info to bet on military strikes on Polymarket
• Sharps vs Squares (22:00) - Chamath explains prediction markets will have sharps with inside information eating squares; draws parallel to Buffett’s returns dropping to market average after Reg FD eliminated info asymmetry
• Liquidity Conference Announced (28:44) - All-In hosting new conference in Yountville wine country May 31-June 3 for capital allocators, LPs, GPs; shutting down French Laundry and all of Yountville
• CBO Death Spiral (32:48) - 2026 deficit at $1.9T (6% of GDP); debt growing from $31T to $56T by 2036; Social Security trust fund runs out 2032; debt-to-GDP at 120% heading to 135%
• State Pension Time Bomb (35:50) - Friedberg warns California alone has nearly $1 trillion in unfunded pension obligations; if Democrats win 2028, these will likely be federalized, piling onto the CBO projections
• CBO’s Low Growth Assumptions (38:00) - Sacks notes CBO projects only 2.2% GDP growth when Q3 2025 was over 4% and Q4 over 5%; believes AI-driven growth could be the way out of the debt spiral
• Freeze Spending Strategy (39:00) - Sacks proposes simply freezing federal spending until the economy grows enough that spending drops to 20% of GDP, currently at 23%
• Federal Employment at 1966 Levels (40:00) - During Trump’s second term, federal employees dropped from 3 million to 2.7 million, lowest since 1966; over 300,000 cut
• Debt-to-GDP Historical Context (42:00) - Chamath shows 300 years of data suggesting debt-to-GDP moves in unison across major economies; if everyone is at 200-300%, nothing really changes that much
• Strong January Jobs (48:06) - Economy added 172,000 private sector jobs in January vs 70,000 expected; government shed 42,000; unemployment back down to 4.3%
• Immigration and Wages (52:00) - Jason argues 75% of illegal immigrants come for jobs; enforcement should target employers especially in construction (#1) and hospitality (#2) rather than city raids
• Ferrari’s First EV (1:03:22) - 1,000+ HP, 4 electric motors, 0-60 in under 2.5s, 330-mile range, heaviest Ferrari ever at 5,100 lbs; interior designed with Jony Ive’s team featuring tactile buttons and Mac-inspired screens
• Autonomous Driving Future (1:07:00) - Chamath predicts car culture will decline as FSD and autonomy make self-driving too risky for insurance; Ferraris will become the luxury exception

Mentions

Companies

• OpenAI / ChatGPT (10:20) - Discussed as example of cloud AI leaking proprietary data
• Anthropic / Claude (8:00) - Claude Opus 4.6 mentioned as orchestration model; API costs at $300/day per agent
• Google / Gemini (10:30) - Mentioned as one of the model builders receiving leaked data
• AWS / Bedrock (14:00) - Chamath’s company is a top 20 Bedrock customer; calls the overhead costs too expensive
• CoreWeave (14:30) - Alternative to AWS but pricing issues with spot vs guaranteed
• Nvidia (16:00) - Incentivized to push down token costs
• Groq (16:00) - Mentioned alongside Nvidia for energy density improvements
• Tesla (1:03:22) - Compared to Ferrari’s approach on interior design; FSD discussed extensively
• Ferrari (1:03:22) - First all-electric vehicle revealed; CEO Benedetto Vigna praised
• Polymarket (19:19) - $700M wagered on Super Bowl; Israeli soldiers allegedly used it for insider trading on military strikes
• Kalshi (20:00) - Prediction market platform; CEO discussed insider trading challenges on CNBC
• Slack (8:30) - Jason upgrading to enterprise version to ingest all messages for AI; Chamath notes it was his #4 investment
• Apple (1:03:22) - Jony Ive’s former company; Ferrari interior described as Mac-inspired
• Craft Ventures (18:00) - Sacks’s firm; partner created Lobster Tank, enterprise-secure version of OpenClaw
• Waymo (1:08:00) - Chamath uses Waymo in the Valley; discussed as future of transportation
• Lexus / Toyota (1:10:00) - Lexus LM and Toyota Alphard luxury minivans not available in US; #1 chauffeur cars in China and Middle East

Products & Technologies

• OpenClaw (4:30) - AI agent platform Jason is obsessed with; investing in 10 startups at $125K each
• FSD (Full Self-Driving) (1:07:00) - Sacks says FSD made him a driver again; Chamath predicts it will end car culture
• Lobster Tank (18:00) - Enterprise-secure version of OpenClaw created by Sacks’s partner at Craft
• Kimmy (8:30) - Local AI model running on Mac Studios for free; used for easy jobs
• Mac Studio (8:00) - Used to run local LLMs with 512GB RAM; Jason has multiple running AI agents

People

• Jason Calacanis (0:08) - Host; running AI agents at his firm; investing in 10 OpenClaw startups
• David Sacks (1:08) - Predicted AI increases knowledge worker demand; argues we’re entering a golden age
• David Friedberg (32:48) - “Dr. Doom” on debt; warns about state pension federalization as biggest risk
• Chamath Palihapitiya (10:20) - Raises on-prem vs cloud question; provides historical debt-to-GDP analysis; curating Liquidity conference
• Matt Schumer (2:30) - Wrote viral article “Something Big Is Happening” about AI career opportunities
• Jensen Huang (1:50) - Referenced for concept of task-based vs purpose-based jobs
• Jony Ive (1:03:22) - Former Apple design chief; worked with partner Marc Newson on Ferrari EV interior
• Marc Newson (1:03:22) - Jony Ive’s partner; also designed the iconic Ford 021C concept car
• Warren Buffett (23:40) - Returns dropped to market average after Reg FD eliminated information asymmetry
• Scott Bessent (33:00) - Treasury Secretary; set 3% deficit-to-GDP target
• Kevin Warsh (46:00) - Potential Fed chair; Friedberg questions what his tightening would do to rates and spending
• Donald Trump (40:00) - Discussed re: federal employment cuts, border policy, potential minimum wage increase
• Benedetto Vigna (1:07:00) - Ferrari CEO praised as talented executive
• Elon Musk (1:07:30) - Model X discussed; wished he had made a minivan or three-row SUV
• Benny Ofir (12:00) - Building Slackbot AI tools that Jason is consulting with

Surprising Quotes

“We hit $300 a day per agent using the Claude API. That’s $100,000 a year per agent. When do tokens outpace the salary of the employee? Because you’re about to hit it.” — Chamath Palihapitiya, 16:40

“There is no attorney-client privilege once you start to use those cloud tools. All of that stuff is complete public domain material.” — Chamath Palihapitiya, 11:00

“The single best investor in the world generated zero alpha once information asymmetry was eliminated. Markets thrive when there’s asymmetry. Billions will be made in asymmetry.” — Chamath Palihapitiya, 23:40

“We’ll look back on this time period as the beginning of a new golden age. But if you remember what politics were like in the late 90s, all anyone talked about was whether Bill Clinton got a blowjob from Lewinsky.” — David Sacks, 55:40

“If I’m GEICO, do I want to have all of my actuaries using all of our proprietary, private, and confidential data on risk pricing in an open instance of an LLM? The answer is no. That’s obvious.” — Chamath Palihapitiya, 13:00

Transcript

0:00 All right, everybody. Welcome back to the number one podcast in the world, the All-In Podcast. With me again, the core four, the original quartet, David Sacks, David Friedberg, Chamath Palihapitiya. I’m Jason Calacanis, and we have a very full docket today.

0:18 All right, topic one, gentlemen. AI acceleration. It was a big week for AI. New study published on Monday, February 9th, in the HBR, Harvard Business Review, suggesting that AI tools intensify work but do not reduce it. Two UC Berkeley researchers spent eight months embedded at a 200 person tech company. What they found, employees who use AI worked at a faster pace, took a broader scope of tasks, and extended work into more hours of the day. Workers reported feeling more productive, but they also felt a little more stress and burnout. Sacks, your hot take here.

1:08 All right. Well, a few points here. Number one, as you may recall on the prediction show for this year, my most contrarian belief is that AI would increase demand for knowledge workers, not put them out of business. And I think you see in this UC Berkeley study, the reason why that might be the case is because the employees who use these tools, they worked faster, took on a broader scope of tasks, actually ended up working more hours in the day. So they did more work, not less, and even more effort rather than less. Not because they were required to, but just because they were more motivated. Their work was getting upleveled. They’re able to offload more menial tasks to AI and it made their work more purposeful and meaningful. So I think we’re moving from what some people like Jensen has called task-based jobs to purpose-based jobs. A key skill of employees is going to be the ability to structure work for themselves and their AI agents.

2:30 That brings me to point number two, which is that there’s a tremendous opportunity this year for employees who are early adopters of these tools, AI natives, to demonstrate their value. They’re going to be the people in meetings who can take an assignment that would have taken days before and get it done in 2 hours. There was an article that went viral this week by Matt Schumer called “Something Big Is Happening” about this career opportunity for AI early adopters.

3:15 Point three, I think you’re going to see massive enterprise adoption of AI this year, not just chatbots but agents. It’s going to be driven by the bottom up. These early adopter employees bringing in consumerized AI tools as opposed to top-down initiatives where the CEO has tasked a team with figuring out how to use AI. Those initiatives are going to take months doing RFPs. While those things are trudging along, there’s going to be early adopter employees who just make the transformation fait accompli. In the same way that consumerized SaaS tools spread from the bottom up in enterprises, I think consumerized AI tools will spread from the bottom up.

4:30 Couldn’t concur or agree more. I did a tweet that got 2 million views. Basically, if you got laid off by Amazon or Microsoft over the last two years, just learn OpenClaw and automate your previous job. Show you know how to use these tools. Every startup I know is hiring for this position, which is somebody who knows how to build and manage agents. There is no job rec for this yet or a title. We used to call it prompt engineer. It’s no longer just prompt engineering. It’s managing and educating and offloading work to an agent and then making sure they’re actually doing it.

5:30 Right now the people in my organization, four of them who are focused on this out of 20, I would say their leverage is between 10 and 20x the other 16. We now have AI going through podcasts looking for the best moments, clipping them, putting them in Google Drive. We have it looking at our YouTube stats, Instagram, TikTok stats and then giving us strategies for how to make clips go more viral. It’s eliminating all the reporting work that knowledge workers do.

6:30 Chamath, you have a take on this? I know you’ve deployed the software factory. Last couple of weeks have been pretty big with Claude Opus 4.6 coming out, ChatGPT Codex coming out, the OpenClaw revolution.

7:00 I think there are two open questions. The first question is, is on-prem the new cloud? Which is weird to think that could even be possible, but we’ve spent since migrating everything to cloud because there were economies of scale. The counterpoint is that in the AI revolution, companies will be fighting for their lives. And it’s very much unclear whether it makes sense for a company to allow the natural leakage of their edge and their confidential and proprietary information out into the wild versus the control they would get if they ran on prem.

8:00 Once you use these tools, it is very difficult for a company to control how their data is used subsequently. If you gave someone a PDF of some really important strategy document and they’re interrogating it with one of these models, you’re leaking all of that prompt and response metadata back to ChatGPT, Gemini, Claude. There’s nothing a company can do about that. If you’re using agents, all those agent traces are going back to these model builders.

9:00 Number two, there was this really interesting ruling around what happens inside these cloud environments, a judge saying there is no attorney-client privilege, confirming that once you start to use those tools, all of that stuff is complete public domain material. If you put these two things together, it creates a very interesting set of questions for enterprises. You will need AI to survive. But if you use the tools as they exist today at a public endpoint, you will give up all control, all security, all confidentiality. The only solution is to have private provision networks, which increases cost, but then if you save a bunch of money because of AI, maybe it all balances out.

10:20 Good insights there. Friedberg, your thoughts on this moment in time when we have people saying it’s happening faster and it’s become recursive? Well, the thinking historically was that it was going to be about recursive model development where we were going to continuously improve the actual model. It may be the case that the output is what’s recursive and that turns out is having the effect everyone was waiting for. A lot of computer scientists who have worked in AI for some time were surprised about this moment.

11:00 We have now seen that every week 5 to 10% of the work we’re doing inside of our venture firm is being moved over to OpenClaw. We call them replicants. We give them a Notion account, a Slack account, Google Docs account. They have their own email. All of this technology was here all along, maybe for the last 6 months. But no company would give the keys to the kingdom to allow these agents to actually act on your behalf because they don’t want to be responsible if it ships your Bitcoin keys to somebody else. In order to use these, you have to trust them. And if you trust them and then you are monitoring them, the results are unbelievable.

12:00 We have also fired up Mac Studios with Kimmy on them. We are moving all the work onto these and then they’ll use Kimmy for most of their easy jobs, which is free. Then they will use Claude 4.6 Opus to orchestrate things. We’ve also created OpenClaw Ultron which is one meta-replicant that is managing the other four and it checks their work. It talks to them all day long about what they’re doing and then summarizes it.

13:00 I was able to take about 30% of my Athena assistant’s work and give it to the replicant. On the average investment team individual, we now have probably 20% of their work being done by agents in real time. And the best part about it is they don’t forget to do work. They don’t make mistakes. Once you put this in, you don’t need to have checklists. They just do it perfectly every single time. And they work.

14:00 I’m upgrading to the highest level of enterprise Slack. Chamath, that’s probably your number four investment in your career. What an amazing investment. I’m ingesting every single Slack message and giving the API key for every single email in our organization to Ultron. They will know everything going on in the organization. It is mind-blowing how fast this is going.

15:00 Just a plug, I’m investing in 10 startups in OpenClaw space, $125K each to come to the accelerator. This is the 100% focus of our firm. When do you guys think enterprises have a huge freakout around all of this and say, “Wow, we’re leaking all of our most important information”?

15:30 I think there’s a big opportunity to take something like OpenClaw and make it enterprise-grade and secure. One of my partners at Craft actually created a new tool called Lobster Tank, which is a version of OpenClaw with enterprise security wrapped around it. On-prem is back. It’s going to happen.

16:00 If I’m GEICO, do I want to have all of my actuaries using all of our proprietary, private, and confidential data on risk pricing in an open instance of an LLM? The answer is no. So now the question is how do you adapt to that? How do you actually generate tokens in that situation? That is a very expensive technical problem. Remember VAX terminals? You could see a resurgence of that idea. A centralized computer and a bunch of dumb terminals with a CLI.

16:40 We with our agents hit $300 a day per agent using the Claude API like instantly and that was doing maybe 10 or 20%. That’s $100,000 a year per agent. We’re getting to a place where we have to say what is the token budget for our best devs. You can clearly see a trend where they need to be at least 2x as productive. This is a very interesting trend. When do tokens outpace the salary of the employee? Because you’re about to hit it. I think superstar developers are already there.

17:30 The rank and file is probably 10-20% max. The trend is what matters. Unless we have some gigantic leap forward in generating output tokens at one-tenth the cost, which I suspect we will have. Nvidia, Groq, Google, AMD are all incentivized to massively push down the token cost. That’s going to happen, but it doesn’t change the trend on confidentiality.

19:19 Let’s talk about prediction markets, gentlemen. They hit critical mass this past weekend at the Super Bowl. More than a billion bet on Kalshi, 700 million on Polymarket, almost two billion dollars in wagering. The media has been obsessing about market manipulation and insider trading. A day-old anonymous Polymarket account correctly predicted 17 out of 20 halftime show bets, including special appearances by Lady Gaga and Ricky Martin, but only profited $17K. Another account created less than 24 hours before the game correctly bet on Bad Bunny’s set list.

20:30 Wall Street Journal this morning: “Israeli soldiers accused of using Polymarket to bet on strikes.” Israel arrested several people, including Army Reserve, for allegedly using classified information to place bets on Israeli military operations. The account Rico Suave 666 traded in more than $150,000 in winnings before going dormant for 6 months.

21:00 The question is, is it really insider trading? If you and I were making a side bet and I knew something, should the government regulate that? This goes back to securities regulation and the concept of insider information. There’s a chart showing the distribution of accounts. A few accounts have a huge amount of money and make almost all the profits, and a lot of accounts have very little money and get burnt through very quickly. The accounts with lots of money generally only trade in things where they have an edge.

22:30 Let’s define some terms. In betting there are sharps who know what’s going to happen and squares which is everybody else, grist for the mill. The thing with prediction markets is there are going to be fundamental markets that are purely about inside information. What can a regulatory body do about that? I think the answer is not much. If you try to regulate this, it looks like a securities market. These things are too fluid, too dynamic, too ephemeral to be legislated like a security.

23:40 In 2000, we introduced Reg FD. If you’re a CFO, you cannot tell an individual stock manager something you don’t tell everybody else. Here is Warren Buffett’s returns pre and post Reg FD. His returns were double the market returns when this kind of information sharing was legal. The minute it became illegal and you had to act on the same edge as everybody else, his returns went to the market return. He generated zero alpha. This is the single best investor in the world. Markets thrive when there’s asymmetry.

25:00 There are a certain percentage of prediction markets about the well functioning of society and the use of inside information gets to the truth faster and I think that has value. If it uncovers corruption or misdeeds, and that’s the incentive it takes for folks to work around what would otherwise be whistleblower laws, that probably benefits society. But there are other markets where some people will set up a market they know about and can control. Unfortunately, there’s no way to discern when a market gets created whether it’s one or the other.

28:44 We’re really excited about a new event. The besties are hosting a new conference, a retreat, a summit in wine country May 31st through June 3rd. It’s called Liquidity. This is for capital allocators, LPs, and GPs. We will convene the best investors in public markets, the best hedge fund managers, the best private market investors, the best growth investors, the best credit investors, and the largest cohort of LPs representing trillions of dollars of capital, and the CEOs of the fastest growing companies in technology.

30:00 We’ll have some presentations, best ideas, build relationships. We’re going to shut down all of Yountville. We’re going to shut down the French Laundry. We’re going to shut down all of it. And it’ll be ours for a two-day playground where we will build relationships, allocate capital, and maybe make some money as a result. Apply at allin.com/events. This is not a general admission type event. It is specifically for capital allocators.

32:48 Okay, the new CBO report is out. Friedberg, you said we are in a debt death spiral. The Congressional Budget Office released its long-term budget forecast Wednesday, February 11th. 2026 deficit is $1.9 trillion, nearly 6% of GDP, much higher than the 3% GDP target from Scott Bessent. Social Security Trust runs out in 2032, one year earlier than previously expected. Debt will grow from $31 trillion today to $56 trillion in 2036. Average of $2.5 trillion per year. Currently at 120% debt to GDP, heading to 135%. For comparison, Japan is 237, Singapore 176, Venezuela 164, Greece 154, UK 94. Twenty years ago, our debt to GDP was but 60%.

34:00 There’s no outlook to 3% deficit to GDP. One of the key assumptions is that the short-term interest rate is modeled to be around 3.1%. But if rates climb closer to 5%, it adds another $650 billion a year of interest expense, which takes interest expense almost up to $2 trillion a year just paying interest on past debt. Because we’re running a deficit, that new interest expense increases the debt every year. So the debt goes up and up just by adding interest on past debt. That becomes the death spiral.

35:50 The trigger point I’m getting more concerned about: if Democrats win the midterms and you end up with a Democrat in the White House in 2028, there’s a bigger problem which is state and local obligations. Social Security is going to run out of money in a few years. They’re going to need to print a lot more money to fund it. California has nearly a trillion dollars of unfunded pension obligations to its public employees. If you end up with a Democrat-controlled House and Democrat president in 2028, you’ll very likely see a federalization of that obligation. When you add that liability to this CBO report, which doesn’t include any of that, that could be not just the straw that breaks the camel’s back, but the concrete that breaks the camel’s back.

38:00 We all agree about the problem of federal spending and the deficit. With respect to the CBO study, however, one of the key assumptions is that CBO projects real GDP will only grow by 2.2% this year. That’s a very low assumption given that we grew by over 4% in Q3 last year and the preliminary number for Q4 was over 5%. Our predictions for GDP growth this year were 5% plus. They predict it slows to 1.8% after 2026. If you believe all this capex in AI infrastructure is going to have a payoff, the growth rates could be a lot higher. Ultimately the way to get out of the debt spiral is strong growth.

39:30 The two key charts are federal net outlays as percent of GDP and federal receipts as percent of GDP. You don’t want those lines more than 3% apart. Historically, tax receipts have bounced around 17%. Federal net outlays have bounced around 20%. Before COVID, spending was around 20%. It bounced to 30% in 2020. Now it’s trending around 23%. If it was up to me, I would just freeze federal spending until the economy grew to the point where spending is 20% of GDP. We’re not even talking about cuts, just limiting the rate of growth until the economy catches up.

40:30 The level of federal employment is at the lowest level since 1966. During Trump’s second term, we’ve gone from roughly 3 million federal employees to a little bit under 2.7 million. Over 300,000 federal employees have been cut. I think that is a good start. When people move from the government workforce into the private workforce, they become productive. They’re making things that grow the economy. And theoretically, they should also make more money. So this is positive from an economic point of view.

42:00 I think you have to take a broader historical context. Does debt to GDP matter? I would say it doesn’t matter. This is a historical look back from about 300 years. What you see is the trend since 1700 to now is up and to the right. And the key observation is that it moves in unison. These things are relative problems. So if the entire world moves in unison, there is an argument that you could end up at 300, 250, 200% of debt to GDP. But if everybody is there, nothing really changes that much.

43:30 The real question is if one country is able to decouple itself and its economic output is so meaningfully different. Is it important to contain the debt? Absolutely. But for practical reasons that impact your daily lived life: earnings, inflation. We know President Trump was elected on a massive mandate to secure the border and look at waste, fraud, and abuse. He drafted the most important and prolific private businessman in the history of the world to be his tip of the spear. They identified hundreds of billions of dollars. But when Congress had to act to solidify these cuts, they haven’t done much of anything.

45:00 If the most conservative Congress in the history of the United States has not done much, then as Friedberg said, it’ll only get worse if there’s ever a Democratic House and control. We have to acknowledge that if debt to GDP moves in unison, the music isn’t up for a very long time. You got to find ways of hedging and owning real durable assets because the underlying currency will fluctuate wildly and fall off a cliff. Things like gold will do much better over time. Debt to GDP will trend into the 2, 3, 4, 500s for all countries because governments are addicted to spending.

46:30 There’s also a question of what Fed action will do. If Kevin Warsh really does want to tighten the Fed’s balance sheet and the Fed slows down buying treasuries, Treasury yields will go up. If that lasts and you go from 3 to 4 to 5% on the short end of the yield curve, it starts to become way too expensive to fund this level of deficit spending because interest expense will just climb and eat it all up.

48:06 We are in a consumer-driven economy and the employment rate is absolutely fantastic. Job openings still at 7 million. Unemployment rate at historical lows for our lifetime. Employment participation rate still at 62%. We still have people who could be participating. All of these problems will be solved if more people were to participate and take those jobs.

49:30 Why don’t they take those jobs? Sometimes it’s a geographic mismatch. Sometimes it’s a skills mismatch. But very often the jobs are not paying enough. Crazy prediction: what if Trump decides he’s going to raise the minimum wage? Not saying I endorse this, but it’s incredibly low at seven bucks an hour. What if Trump said we’re going to add a dollar or $2 to it over each year of the next three? This would be incredibly popular and might get people off the sidelines.

50:30 Sacks pushes back: the economic literature shows raising the minimum wage makes it illegal to hire someone whose labor is worth less than the minimum wage. It creates higher unemployment in those segments. If the minimum wage were a panacea, why wouldn’t you make it $100 an hour? Jason counters: looking at what happened in Australia, Scandinavia, Seattle, San Francisco, they have higher minimum wages and higher happiness. Businesses and prices go up about 10 to 20%.

52:00 I got to say on all this economic data, we are at the beginning of an economic boom. We saw GDP growth rates in Q3 and Q4 last year over 4% and over 5%. We just had a January job report where the economy added 172,000 new private sector jobs. This blew away the expectation of around 70,000. Government shed 42,000 jobs. Unemployment back down to 4.3%.

53:00 In Trump’s second term, 615,000 new private sector jobs have been created while over 300,000 government jobs have been cut. Construction, especially non-residential construction related to data centers and AI boom, added 33,000 new jobs in January. The capex for this year from the four leading hyperscalers is $600 billion. That’s a roughly 2% tailwind to GDP growth just from four companies. That doesn’t include all the ROI from that infrastructure.

54:30 We have a boom going on and everyone’s kind of blackpilling about this. They’re focusing on this CBO report with unrealistically low growth rates. When we look back on this period, it could end up being a little bit like the late 90s. We had phenomenal economic growth, the internet. But all anyone talked about was whether Bill Clinton got a blowjob from Lewinsky. I suspect we’ll look back on this time period as the beginning of a new golden age. I agree.

55:30 Anytime a statistic is 10-15%, I highlight it. Unemployment went from 4.1% where Trump inherited it up to 4.5, about a 10% increase. But to your point it’s gone down because the border is closed and we’ve got good news in the economy. All the job creation has been enjoyed by native born Americans. All the job loss has been on non-native-born Americans. The slight tick up in October was because of the October 1st DOGE buyouts. All voluntary. Now unemployment has ticked back down. Job creation right now is strong.

57:00 The top two areas where illegal aliens are working: construction number one, and leisure and hospitality number two. Two and a half million people working in those categories. If you want to see more Americans take jobs and wages go up, go to those businesses and fine them for hiring illegal aliens. You go to the construction site. Everybody checks in. They take pictures. Then they go to the business owner and say, “Show us these people’s paycheck stubs.” It’s super simple. This has been done for decades.

58:30 Here’s a very important case from 2017. The Justice Department and ICE went after a company hiring illegal aliens. 95 million recovered, the largest ever levied immigration case. We can solve almost all of the immigration issues by doing basic surveillance and detective work, asking businesses to show paycheck stubs. According to the LA Times survey, 75% of immigrants come here for better job opportunities. They are not coming to commit crimes or get benefits. The blame needs to be on the business owners who are creating the incentive.

1:00:30 Your suggestion is to do this for every company in America? No, I’m being specific. You pick the number one employer of illegal aliens, 2.5 million people working construction. You start with the largest construction sites and then work backwards. Then the largest restaurant and hotel chains. Steven Miller should stop doing the crazy raids and should go and surveil construction sites. This is well within ICE’s purview. They’ve been doing this for 30 years.

1:02:00 Both problems are equally important. The gang bangers are a small number. The people working in construction and hotels are a big number. The thing we’re not doing at scale is going after the businesses that are creating the incentive for the majority of people who come here.

1:03:22 Ferrari has a new car coming out. It’s going to be their first all-electric vehicle. Very polarizing. 1,000 plus horsepower, four electric motors, 0 to 60 in under 2.5 seconds, 330-mile range. It’s the heaviest Ferrari ever, 5,100 pounds compared to the iconic F40 which was but 3,000 pounds. It’s going to launch in May 2026. We got to see the interior and this is what everybody’s buzzing about. It’s gone viral.

1:04:00 Former Apple design chief Jony Ive and his partner Marc Newson who designed the iconic Ford 021C concept car were involved in this. You have this beautiful square glass key like an iPhone. You put it in and the yellow Ferrari yellow drains out and goes into the shifter. The screen looks very Mac-inspired except unlike Tesla which is no buttons, they’re adding buttons here and making them very tactile. All the sports car enthusiasts love tactile memory-based buttons. Finally, turning the car on is like starting up a jet. You have a launch button you twist and press and it makes the whole car turn Ferrari red.

1:05:00 Sacks on the interior: I thought it found a compromise between the all-glass cockpit of a Tesla versus a totally analog old Ferrari interior. Nice balance between the simplicity of that iPad screen but also having enough buttons that you develop muscle memory. But I hate the look of the outside of this car. This to me looks like a Corvette or even like a Pontiac Firebird Trans Am. It looks like a Model 3. The front grill looks terrible. A Ferrari should look swoopier, curvier.

1:06:30 Chamath, you a buyer? I’ve had a Ferrari. There’s just something that’s very unique. There’s a Ferrari experience different from every other car. The new CEO, Benedetto Vigna, is a very talented executive. But we are racing against time. FSD and autonomy is going to shift the number of people that even know what it means to drive. It will feel like when we look at somebody who embraces thoroughbred racing. The risk will not make any sense for most people under most conditions.

1:07:30 The car culture in America was a profound part of the American culture. Driving from A to B on vacation, the sense of freedom, the interstate highway system. These were huge parts of what made America great. And now I think it’s all going to change. In places like China and India, they’re always going to have a market. But in the United States, it’s going to become so expensive to pay for insurance if you are driving yourself. Ferraris will be the luxury exception where you pay for the experience, the insurance, all of it. The rest of us will be using FSD or Waymo.

1:09:00 We have two Model Ys and need another car. I’m so mad about the X being depreciated. I have five kids so the X is the only car that can manage seven people. I wish Elon would have made the minivan or three-row SUV. When I was in Abu Dhabi, I saw my dream car. The Lexus LM. The doors open and it’s like first-class airline seats. The front is completely blacked out for total privacy. It’s got beautiful captain’s chairs, a full monitor. These are like Etihad first-class airline seats. These are the number one cars in China, Singapore, the Middle East for chauffeur-driven cars. None of these are available in the US.

1:11:00 All right boys, I love you very much. That’s another amazing episode of the All-In Podcast. Episode 261, 261 weeks and counting. Strong.