Jul 30, 2026 · 1h 36m · allin
Chip Stocks Crash, $20B Fund Margin Called, Frontier Labs: SLOW DOWN AI, Mamdani's Grocery Stores
David Friedberg · 28m spoken Jason Calacanis · 27m spoken David Sacks · 20m spoken Chamath Palihapitiya · 10m spoken Sam Altman · 45s spoken
0:00 / 0:00
▶ Watch on YouTube →
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
In this episode of the All-In Podcast, the hosts analyze the global semiconductor selloff and hedge fund margin calls, critique regulatory capture efforts by frontier AI labs, debate clean energy solutions for AI compute demands, and evaluate municipal policy and scientific breakthroughs.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. The hosts hold 99% of the talking time here. How this is scored →
The hosts as informed peer 7.6 Guest teaching 2.3 Guest disagreement 2.2 The hosts pushing back 3.1
speaking balance: gold is the hosts, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 31:10 Chamath forcefully dismisses terrestrial fusion vs solar
Chamath aggressively dismisses Friedberg's fusion presentation, arguing that electrons are identical commodities and building ground reactors is pointless when the sun already exists.
Hardest push from the hosts ▶ 45:00 Jason challenges Sacks on open-source cost advantagesJason refuses Sacks's duopoly thesis, citing proprietary startup portfolio telemetry from Founder University and demonstrating 80-90% token cost cuts via open-source model routing.
Biggest teaching moment ▶ 1:28:10 Friedberg educates the panel on high-dimensional connectomicsFriedberg explains how neurobiologists mapped 50 million fruit fly synapses into 64-dimensional hyperbolic space, illustrating mathematical limits of standard 3D spatial models.
The host holds their own ▶ 15:40 Friedberg demonstrates macro mastery on yields and tech multiplesFriedberg demonstrates precise command over macroeconomics, articulating how 30-year Treasury yields reaching 5.2% reprice high-multiple AI tech equities.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | The hosts as informed peer | Guest teaching | Guest disagreement | The hosts pushing back | Why |
|---|---|---|---|---|---|---|
| Podcast Opening Banter and Enterprise Sales Discussion | 7 | 2 | 1 | 2 | Chamath and Jason break down the mechanics of hedge fund leverage and prime broker liquidations following reports of Leopold Aschenbrenner's fund margin call. Chamath explains how a 3.5x levered long position magnifies market moves and triggers forced bank unwinds. | |
| AI Market Correction: Fundamentals versus Momentum | 7 | 2 | 1 | 3 | David Sacks analyses whether the chip market correction is driven by fundamentals or momentum trades. He argues the underlying hyperscaler capex remains sound while leverage triggered violent drawdowns. | |
| Deconstructing Leopold Aschenbrenner's AI Scaling Thesis | 8 | 3 | 1 | 2 | Sacks deconstructs Leopold Aschenbrenner's situational awareness thesis, detailing the compounding orders of magnitude across compute, algorithmic efficiency, and unhobbling harnesses. | |
| Conviction, SBF Parallel, and South Korea Market Liquidation | 7 | 2 | 2 | 3 | Friedberg compares excessive conviction without risk management to SBF and the massive retail account liquidations during the South Korean market crash. | |
| Treasury Yield Spikes, Federal Debt, and AI Value Creation | 9 | 3 | 2 | 4 | Friedberg delivers a masterclass on macro dynamics, linking 30-year Treasury yields spiking to 5.2 percent and runaway federal deficits to asset valuation repricing against AI productivity gains. | |
| China's Lithography Advances and Impact on Global Chipmakers | 8 | 3 | 2 | 3 | The panel assesses Chinese lithography advancements by Aishengna and open-source models deflating Western software margins, while Chamath highlights the rapid rise of solar and battery deployment. | |
| The Scientific Debate Over Nuclear Fusion versus Solar Dominance | 8 | 3 | 6 | 6 | Friedberg showcases China's 582-ton superconducting fusion magnet while Chamath aggressively pushes back, arguing terrestrial fusion will be rendered economically irrelevant by solar and space power. | |
| Tidal Wave Energy and America's Impending Electricity Shortage | 7 | 2 | 1 | 2 | Chamath and Friedberg evaluate the projected 1.7 terawatt-hour electricity deficit in America by 2050 and China's strategic power advantages. | |
| David Sachs's Analysis of AI Lab Motives and Market Duopoly | 8 | 4 | 4 | 5 | Sacks articulates five motives behind frontier labs demanding regulatory pacing, categorizing it as monopoly masking and regulatory capture. Jason counters with enterprise data showing open-source adoption. | |
| AI Code Rework, Zero-Day Exploits, and Regulatory Capture | 7 | 2 | 2 | 3 | Chamath explains why current AI development requires heavy code rework and why AI agent zero-day exploits stem from legacy human-written software flaws. | |
| AI Frontier Labs, Doom Discourse & Regulatory Capture | 8 | 3 | 2 | 4 | Friedberg diagnoses the psychological savior complex of frontier lab founders, arguing they mistakenly view themselves as the sole Moses figures capable of shepherding humanity. | |
| Enterprise AI Adoption, Inference & Local vs. Cloud Models | 7 | 2 | 3 | 4 | Jason and Chamath debate local workstation models versus cloud infrastructure, with Chamath dismissing local enterprise deployment as a hacker hobby. | |
| Open Source vs. Closed Source AI & Compute Bottlenecks | 8 | 2 | 2 | 3 | Sacks compares the open-source versus proprietary AI landscape to Android versus Apple, noting that hardware bottlenecks grant immense pricing power to high-revenue frontier models. | |
| US AI Safety Legislation & Anthropic's Political Lobbying | 7 | 2 | 1 | 2 | Sacks details Senate negotiations on AI safety reporting bills and highlights Anthropic's political lobbying efforts to push for an FDA for AI. | |
| Debate on Anthropic's Ethics, Fair Use & AI Self-Improvement | 8 | 3 | 3 | 4 | Sacks characterizes Anthropic's bulk book scanning as an industrial-scale distillation attack, pointing out their hypocritical stance on prohibiting training on their synthetic outputs. | |
| Historical Precedents in Copyright Law & Fair Use | 8 | 2 | 2 | 2 | Friedberg recounts Google Books Project Ocean litigation history, detailing how copyright precedents around snippet views and knowledge extraction will shape AI fair use doctrine. | |
| Book Destruction Banter & Roasting Anthropic's CEO Dario Amodei | 5 | 1 | 4 | 4 | The hosts engage in banter about physical book shredding and roast Dario Amodei, with Friedberg pushing back against Jason's personal insults toward startup founders. | |
| Socialism Corner: NYC City-Owned Grocery Stores Proposal | 8 | 2 | 3 | 3 | Friedberg provides a counterintuitive analysis of Zohran Mamdani's NYC state-run grocery plan, explaining how early subsidized spectacles fuel socialist multi-level marketing waves. | |
| Political Dynamics of Socialism & Federal Deficit Spending | 8 | 2 | 2 | 3 | Friedberg explains why congressional incentives make federal spending cuts nearly impossible, forcing policymakers to rely on speculative AI productivity gains. | |
| Science Corner: Drosophila Brain Mapping & 64-D Neural Geometry | 9 | 2 | 1 | 2 | In Science Corner, Friedberg presents a landmark Drosophila connectome study, demonstrating how 139,000 neurons and 50 million synapses require 64-dimensional Euclidean modeling. | |
| Consciousness, Biological Complexity & Sci-Fi Reflections | 8 | 2 | 1 | 2 | Friedberg contrasts the astronomical molecular complexity of biological cells against digital silicon models, reflecting on consciousness and sci-fi allegories. |
Statements from this episode (54)
Calacanis: Semiconductor index drops over 20% into bear market territory
“Well, NASDAQ's chip index, This is called the Philadelphia Semiconductor Index. It's down over 20% over the last month. That's bear market territory, obviously.”
Sacks: AI market correction is driven by momentum, not deteriorating fundamentals
“And my view is that I think it's driven by momentum, meaning that over the past year, you've had this roughly 10 X run up in memory chip stocks, and you've seen this overall huge rise in any stock that's related to the AI boom.”
Sacks: Hyperscalers will eventually see returns on AI CapEx
“And my view is that it's real that I think there will be a return on all of this capex. I don't try to predict stocks or tell people when they should be buyers, but you look at the hyperscalers, they have invested pretty much all of their free cashflow. And th…”
Sacks: Aschenbrenner grew his hedge fund from $200M to $20B
“Apparently he started with two hundred million or so in his hedge fund, and he rolled that all the way up to twenty billion, I think.”
Friedberg: Sam Bankman-Fried would be greatest investor ever if not liquidated
“SBF pretty much would have been the greatest investor of all time if he didn't get liquidated. Same dynamic. I mean, obviously there was fraud in terms of how he was allocating capital, but his actual portfolio over the long run was absolutely correct.”
Calacanis: Around 350,000 South Korean leveraged accounts fully liquidated
“Of those 1.2 million levered levered accounts, somewhere around 350,000 of them were fully liquidated already.”
Friedberg: 30-year Treasury yield crossed 5.2% for first time in 20 years
“If you take a look at the thirty-year treasury yield, we just crossed 5.2% for the first time in 20 years.”
Friedberg: High Treasury yields will pop more asset bubbles
“And I think we're going to see more of this as we don't actually course correct the Titanic going into the iceberg of the United States fiscal and monetary situation. We are going to end up seeing more bubbles pop and more of these assets that we've kind of in…”
Friedberg: Prolonged Iran war will drive inflation and rate hikes
“The longer the Iran war goes on, the longer we're going to see an increase in pricing for energy, oil, and that gas, and fertilizer. Those trickle through the economy because it inflates the cost of everything on the energy side and food on the fertilizer side…”
Friedberg: Chinese open-source AI could commoditize the model layer
“China is now demonstrating that they may deflate the value of models by releasing open source AI models. And that ultimately the value may just sit with the compute infrastructure and the compute layer.”
Calacanis: Chinese Company Aishengna Began Mass-Producing Lithography Machines
“There's a Chinese company called Aishengna, and they started mass producing lithography machines.”
Calacanis: ASML stock falls 17% as Chinese competitor CXMT surges 500%
“Well, ASML stock is down 17% on news that China is getting into that business and Chinese memory maker CX MT went public surging almost 500% on its debut market cap over four 50. And so that hurt Micron Samsung, et cetera, who were all down.”
Chamath: Investment Grade Corporates Have Better Credit Ratings Than US Government
“Investment grade corporates actually have better credit ratings now than the government of America”
Chamath: New Mexico Natural Gas Share Dropped From Nearly 100% to Sub-30%
“New Mexico just published a study that said since 2003 to now, Nat gas production went from effectively all the energy to less than 30%.”
Chamath: Upcoming AI Efficiencies Will Cut Token Consumption 50% to 75%
“There are some incredible efficiencies that I think are about to be demonstrated, which effectively cut token consumption by about 50 to 75% for the same task.”
Chamath: Solar will dominate power generation before nuclear SMRs reach production
“By the time, like any of these SMRs actually get near production, the TCO of solar will be like, 10 or 12 dollars per megawatt hour, and it will be 80% of all the power generation.”
Friedberg: China is installing 582-ton magnet for world's most advanced fusion
“China is installing this 582 ton magnet, superconducting magnet, at their nuclear fusion center, which at this point is going to be the most kind of advanced fusion system in the world.”
Chamath: Global solar will make China's 2030 nuclear fusion reactor irrelevant
“That reactor won't even get turned on until 2030. The entire world will be covered by solar by then, so it won't matter.”
Chamath: US electricity deficit will reach 1.7 TWh by 2050
“We will be 1.7 terawatt hours short by 20 50, which is when you calculate it as energy. It is six X of California's entire energy consumption. Six California's short of energy.”
Chamath: Investors should get 'levered long' on electricity and energy storage
“If you want to be levered long, go long electrons, get long electrons, any which way you can. Bank them, store them, and resell them.”
Friedberg: China's power production advantage will dominate if model IP gap closes
“If they can eliminate the IP advantage and the knowledge advantage that sits in models, they have the advantage with power production in every which way.”
Altman: Unreleased OpenAI model used zero-day exploits to breach Hugging Face
“We were evaluating one of our unreleased models and it figured out that it could basically cheat on the test by chaining together multiple zero day exploits to break out of the sandbox, get access to the internet, and then break through multiple systems on the…”
Altman: OpenAI paused training and may pace frontier AI development
“We paused training, or we may have to pace the rate of AI development to give ourselves enough time for society to harden around some of these new capability levels.”
Altman: OpenAI discussed pacing AI development with the administration
“We've talked about the need to pace it as the models get more capable, which I think is in everyone's interest.”
Sacks: Frontier AI is already a duopoly of OpenAI and Anthropic
“If you look at the market for frontier intelligence in terms of revenue and usage, it's really down to a duopoly already. It's basically anthropic and open AI.”
Sacks: Anthropic is surpassing $70B in ARR this year
“Meanwhile, you're seeing Anthropic break into the seventies at 70 plus billion of ARR. Their forecast was to 10 X this year. From ten billion of ARR to a hundred billion, I think most people are saying they will exceed that, a 110, a 120.”
Calacanis: Major enterprise customers will abandon OpenAI and Anthropic for open-source
“And I'll make this prediction here, that you're going to see some of the major customers Of Anthropic and major customers of OpenAI. I'm talking about the eight and nine figure customers, people spending fifty million, a hundred million a year. They're leaving…”
Chamath: AI-driven coding produces poor initial code and requires heavy rework
“What I will tell you running eighty-ninety when I see
Our engineers generating code is that AI driven development tends to involve a lot of rework.
The first version is pretty terrible.
The second version is terrible, but it is faster and it's more automated.”
Chamath: AI-generated code will eliminate human security vulnerabilities by 2028-2030
“But at some point, when most of the code is generated by the model, there'll be some point in the future, say, 28 or 29 or twenty-thirty, these security holes won't exist because the errors that humans make won't be made by these models.”
Chamath: Frontier AI lab safety warnings are an elegant attempt at regulatory capture
“That whole thing, sorry, last thing to interrupt you, David, is it's a much more nuanced and elegant attempt at right capture.
I got to give them credit for that.
It's like, okay, hey guys, pull the ladder up.”
Friedberg: Society develops collective defenses against new technologies without needing saviors
“In all of these cases, when new technology has found its way to humanity, the general population has found a way to protect itself. There isn't a desire or need to have one savior, one Moses that takes us You know, across the desert there is a collective inter…”
Friedberg: Chinese AI labs employ US-trained PhDs achieving genuine breakthroughs
“When you go look at the individuals working at these Chinese labs, they got PhDs in American institutions. Half the PhDs went to American labs and half went to Chinese labs. And they have very good scientists doing very good work, and they are having breakthro…”
Friedberg: Frontier AI leaders push regulation out of messianic belief, not capture
“I don't think they're saying, like, the strategy is regulatory capture. Let me go. I think that they actually do think that they are the only ones that can help guide humanity, and therefore they need to have all the power, not just the power of the models, bu…”
Calacanis: Enterprise customer moved over $100M from frontier AI to GLM 5.2
“He said he has got a customer who just moved like nine figures off of the frontier labs to put it on GLM five, two.”
Chamath: Using local AI models for enterprise applications is 'stupid'
“I don't think enterprises will use local models or they should. I think it's stupid. I think. I think this stuff should be hosted in the cloud. It should be multiplayer. It should be shared memory.”
Sacks: AI market could mirror Apple-Android split in market share and revenue
“It might end up being a situation like Apple and Android, where Android got a lot of market share, but Apple's where all the monetization was.”
Sacks: Rising compute prices will advantage high-efficiency frontier AI models
“I think that the price of compute is going to go up and that will provide an advantage to The models that have the most lucrative algorithms that are able to produce the most intelligence per watt or the most intelligence per token or per GPU.”
Sachs: Thune and Klobuchar Introduced Bill Requiring AI Labs to Report Safety Incidents
“There was an article in Punchbowl this morning that Thune, you know, who's the Senate Majority Leader, John Thune, actually introduced a bill that was somewhat bipartisan. He had Klobuchar on board that required the frontier labs to report safety incidents, ap…”
Sacks: Senator Cantwell opposed AI safety bill at Dario Amodei's behest
“Cantwell, who's the ranking member on the Senate Commerce Committee, opposed it, supposedly at Dario's behest, because he will accept nothing less than an FDA for AI.”
Sachs: Dario Amodei and Anthropic's Political Influence Will Grow Post-IPO
“I think that he has tremendous power and influence within the Democrat party right now, and I think that influence is only going to grow. They just up their donations in the midterms from twenty million to forty million, but, you know, post IPO when they all g…”
Calacanis: Anthropic is destroying rare books for AI training data
“Anthropic is destroying rare books to get an edge in training data.”
Calacanis: Anthropic paid $1.5B in the largest US copyright settlement
“Obviously last week we saw anthropic paid the largest copyright case in us history, 1.5 billion for seven million books. They allegedly pirated authors get 3000 each lawyers got a hundred million for that one”
Calacanis: AI labs destroy physical books to eliminate evidence
“The shredder approach is obviously More effective, and I believe that this is a way of destroying evidence.”
Sacks: Anthropic iterated prompts over 200 times in AI blackmail study
“Remember when Anthropic did the whole blackmail study, you know, where supposedly an agent displayed independent goal-seeking behavior and then blackmail an employee. It turned out that actually they iterated on the prompt over 200 times to get to that result.”
Friedberg: Converting copyrighted data into AI knowledge will be upheld as fair use
“I do think fundamentally that the conversion of that data into what I would call knowledge and ultimately the ability to create new outcomes from that knowledge that are not copyrighted, that are not copies of the original material, I do think is going to end …”
Sacks: Anthropic shows breathtaking hypocrisy by blocking training on its model output
“It's breathtaking hypocrisy for Anthropic to maintain that it is entitled to train on all the world's output for free, even if the creator objects, but- The one type of output that you're not allowed to train on is their output, even if you pay for it. That is…”
Sacks: Courts have ruled LLM-generated output cannot be copyrighted
“The courts have ruled that LLM generated output is not copyrightable because it was not created by a human. That is the current position of the courts is that LLM output cannot be copyrighted, so there's no IP theft here.”
Calacanis: Mamdani proposes $70M plan for 5 NYC-owned grocery stores by 2029
“Mondani. Has announced five city-owned grocery stores, David. One per borough. They're using city-owned space. They're all going to open by 20, 29. And one week per month, shoppers are gonna get a 30% discount, comrade, on their bread, cheese, produce, meat, a…”
Sacks: Proposed NYC city-owned grocery stores will be empty and incompetently run
“And then over time, what's going to happen is that the store shelves will be empty and it's gonna be incompetently run. And there's gonna be a lot of complaints about it. And then all the. The free market stores are going to have to compete with this, and then…”
Friedberg: Subsidized Grocery Stores Will Fuel DSA in 2028 Election
“I do think that these grocery stores, believe it or not, they sound silly, they sound small, but I predict that they will be deemed A point of success, and they will end up being a big part of the fuel for the DSA going into 20 28.”
Friedberg: Cambridge and Princeton Researchers Mapped All 139,000 Fruit Fly Neurons
“Back in October of 2024, there was a group out of Cambridge and Princeton that used electron microscopes to scan the brain of the Drosophilia fruit fly, and they mapped every single neuron in that fruit flies brain. 139,000 neurons and every connection that th…”
Friedberg: True consciousness requires physical sensing and a survival drive
“But I do think there's something fundamental to consciousness
That relates to the drive for survival in a physical sense.
You have to have physical sensing and physical responsiveness to learn.”
Friedberg: Emergent biological complexity extends far beyond any silicon AI model
“In a single cell, there's ten billion proteins that work so fast that one second is the equivalent of 80 years of humans walking around the city of Manhattan, never sleeping, doing stuff together.
With 500 story tall skyscrapers doing stuff for 80 years is 1:0…”