All-In PodcastDueling Presidential interviews, SpaceX’s big catch, Robotaxis, Uber buying Expedia?, Nuclear NIMBY
CHAPTERS
- 0:00 – 2:30
Cold Open, Banter, and Election Night Plans
The besties open with light banter, a plug for Freeberg’s Supergut, and an announcement of their November 5th election‑night livestream. They joke about possibly hosting from Mar‑a‑Lago if Trump’s odds look strong, setting up the episode’s focus on 2024 politics and prediction markets.
- •Freeberg plugs Supergut’s nationwide Target launch; joking about commercials in the pod.
- •Announcement of All‑In election‑night livestream on November 5, with Sacks hosting.
- •Speculation about traveling to Mar‑a‑Lago if Trump appears to be cruising to victory.
- •Early hint that prediction markets currently favor Trump.
- 2:30 – 7:00
Polls, Prediction Markets, and a ‘Baked’ Election
The group dives into why betting markets price Trump higher than traditional polling and whether recent high‑profile interviews can move the needle. They argue that most voters are entrenched, and any remaining swing comes down to a narrow band of independents and late‑breaking events.
- •Sachs explains difference between polls (vote share) and prediction markets (probabilities).
- •Nate Silver model sits around 50/50 while Polymarket shows ~60/40 or 65/35 for Trump.
- •Sachs attributes discrepancy to markets reacting to momentum faster than polls.
- •Friedberg questions whether Trump’s Bloomberg and Harris’s Fox interviews can still change minds.
- •Consensus that the race feels largely locked in unless a major October surprise emerges.
- 7:00 – 13:00
Dueling Trump–Harris Interviews and Media Tribalism
They compare Trump’s long Bloomberg interview to Harris’s shorter Bret Baier appearance on Fox, emphasizing how partisan media framed each as a triumph for their own side. Chamath and Sachs critique Harris for non‑answers and failing to define herself relative to Biden, while emphasizing Trump’s comfort in adversarial settings.
- •Friedberg notes left and right media each declared their candidate victorious, reinforcing priors.
- •Sachs cites Bret Baier’s claim Harris’s staff tried to cut the interview short ~26 minutes in.
- •Chamath: Harris was composed but substantively weak, especially on regrets about the last 3.5 years and border policy.
- •Sachs: Harris still can’t articulate how she differs from Biden on policy; fundamental campaign problem.
- •Trump is portrayed as relishing long, combative interviews and winning over live audiences.
- 13:00 – 18:30
JD Vance, Election Legitimacy, and Platform Bias
The discussion shifts to JD Vance’s stance on certifying the 2020 election and his combative media interviews. Friedberg uses a broader lens to talk about systemic biases in media, social platforms, and election rules rather than relitigating 2020, arguing that structural transparency is more important than gotcha questions.
- •Jason presses Sacks and Friedberg about Vance saying he wouldn’t have certified the 2020 results.
- •Sacks dismisses January 6th certification questions as no longer salient to persuadable voters.
- •They praise Vance’s viral exchange with Martha Raddatz about foreign gangs and apartment buildings as more relevant to voters.
- •Friedberg argues both parties have raised concerns over social media manipulation and misinformation in different cycles.
- •He calls for clear, public rules on content moderation and voter verification, rather than dismissing concerns as bad faith.
- •Chamath praises Friedberg’s answer as powerful but too nuanced for the “reductive masses.”
- 18:30 – 21:00
Election Integrity, Voter ID, and Federal vs State Power
They explore reforms that could make future elections ‘above reproach’ and thus reduce election denial. The conversation touches on voter ID, purging voter rolls, DOJ lawsuits, and whether the federal government should set minimum standards without overriding states’ constitutional role.
- •Ideas floated: make Election Day a federal holiday; require voter ID; clean voter rolls.
- •Sachs criticizes Biden’s DOJ for suing Virginia over removing non‑citizens from rolls and California for banning voter ID checks.
- •He argues one‑party states could entrench themselves in federal elections via lax rules.
- •Consensus that clearer, more robust standards would reduce post‑election disputes.
- •Reference to Heritage Foundation’s struggle to prove large‑scale fraud while still arguing for higher integrity standards.
- 21:00 – 25:40
SpaceX Starship’s Booster Catch and the New Space Economics
The hosts marvel at SpaceX’s successful ‘chopsticks’ catch of the 23‑story Super Heavy booster, calling it a historic engineering milestone. Friedberg breaks down how full reusability, cheap propellant, and falling hardware costs could drive launch costs down by orders of magnitude, enabling Mars colonization and a vast new space economy.
- •Visual and emotional reaction to the booster being caught—called a ‘marvel of human ingenuity.’
- •Friedberg explains cost per kilogram to LEO dropping from ~$10,000 to ~$1,000, with a path to ~$10.
- •Starship payload ~150–200 tons; booster holds ~3,400 tons of propellant costing about $1M per launch.
- •Payload estimates: Starship+booster hardware cost may reach ~$35M; with 10 reuses this becomes ~$3.5M/launch plus fuel.
- •At ~$10/kg, half a million tons of material to Mars becomes economically feasible, including ISRU fuel production on Mars.
- 25:40 – 31:40
Starlink Scale, Tesla Robotaxis, and the Robobus Vision
They pivot to Starlink’s subscriber growth and the unveiling of Tesla’s Cybercab robotaxi and Robobus. The group speculates about Starlink becoming the biggest subscription business ever and riffs on how a modular robobus platform could transform transportation, housing, and disaster response.
- •Starlink at ~4M subs paying about $100/month already rivals major ISPs.
- •Speculation: Starlink could reach 100–500M subs, outgrowing Netflix and telcos.
- •Conversation about why we may later see copper wire and legacy ISPs as historical oddities.
- •Tesla’s Cybercab receives praise for design but criticism for lacking steering wheel/pedals—Chamath would buy immediately if drivable.
- •Robobus viewed as a flexible ‘sled’ that could serve as mobile homes, ADUs, Airbnb‑style guest units, or emergency shelter post‑disasters.
- •Shout‑out to Tesla exec Omid Afshar’s promotion; Jason notes bus shuttles to JFK.
- •Sachs reveals he finally bought a Model S Plaid and uses FSD daily.
- 31:40 – 45:19
Uber–Expedia Rumors: Super App Dream vs AI Agent Reality
The Financial Times report that Uber explored acquiring Expedia sparks a sharp strategic debate. Chamath sees OTAs as brittle UIs doomed by AI agents, while Friedberg outlines a financially compelling deal thesis; Sachs questions whether users want vacation‑booking bolted into an app they open for instant rides and food.
- •Uber (≈$170B mkt cap) was rumored to eye Expedia (~$20B), which popped on the news; Uber stock dipped.
- •Chamath calls a $30B‑ish deal “stupid,” arguing Expedia is a front end on commoditized flight/hotel data vulnerable to Perplexity‑style agents executing direct to airlines.
- •Friedberg’s PE‑style model: Expedia EV ~$22B net of cash, current EBITDA ~$3B; with G&A and marketing cuts plus cross‑selling into Uber’s 150M MAUs, EBITDA could reach ~$6B (≈4x EV/EBITDA).
- •Sachs rejects the cross‑sell logic: Uber’s core UX is fast, low‑friction transaction; adding vacation booking is clutter, unlike Uber Eats which is tightly adjacent.
- •Chamath’s macro view: in an ‘agentic’ future, surface‑level UIs lose value; durable assets are data and deep services, not booking front‑ends.
- •Jason suggests limited synergies around hotels/VRBO and potentially rebranding VRBO as ‘Uber Travel’ as a more rational carve‑out.
- 45:19 – 50:40
Nuclear ‘Vibe Shift’ or NIMBY Wall? Tech’s SMR Bet
Big tech’s turn toward nuclear—Amazon, Google, Microsoft engaging SMR projects—triggers a deep argument about whether these announcements represent real capital commitments or mostly conditional PR. Chamath stresses the difference between if‑it‑works offtake agreements and true balance‑sheet risk; Friedberg returns to macro energy constraints.
- •Amazon announces a $500M plan around SMRs near a Virginia plant, tied to $35B in local data center investment by 2040.
- •Google signs nuclear offtake deals (e.g., Kairos Power); Microsoft is reviving part of Three Mile Island.
- •Chamath dissects deal structure: many are contingent offtakes (‘if it works and regulators approve’), not guaranteed capex.
- •He argues real risk capital and regulatory clearance remain the core bottlenecks; tech has not yet put billions of unconditioned equity at stake.
- •Friedberg reiterates that as GDP and AI demand climb, nuclear becomes economically necessary because wind/solar/geothermal alone can’t scale quickly enough, particularly in the US regulatory environment.
- 50:40 – 58:00
Safety Records vs ‘Luxury Beliefs’: The Nuclear NIMBY Showdown
The episode’s most contentious segment pits Friedberg’s data‑driven defense of nuclear safety and modern Gen‑3/4 reactors against Sacks’s visceral NIMBY discomfort and political realism. They clash over casualty stats, meltdown risks, SMR maturity, and whether pro‑nuclear advocacy is an elite ‘luxury belief’ foisted on poorer communities.
- •Sacks: local communities won’t want reactors in their backyard, regardless of AI or competitiveness; calls elite pro‑nuclear stance a ‘luxury belief’.
- •Friedberg: compares anti‑nuclear fear to early opposition to airplanes; argues the risk/benefit ratio overwhelmingly favors nuclear given global energy needs.
- •He cites ~440 reactors in 32 countries, with only three major incidents: Three Mile Island (0 deaths), Fukushima (1 direct death), Chernobyl (~46 direct deaths plus ~15 thyroid cancers), stressing orders‑of‑magnitude lower harm than fossil fuels.
- •Debate over whether statistics include long‑term radiation effects and why exclusion zones still exist.
- •Friedberg distinguishes Gen‑1/2 vs Gen‑3/4 reactors, arguing newer designs are far less prone—or in some SMR designs, structurally unable—to melt down in the same way.
- •Chamath points out many next‑gen SMRs are still theoretical from a commercialization and failure‑mode standpoint; functional examples exist abroad but not yet at scale in the US.
- •They eventually converge on a partial compromise: reactors far from cities (e.g., deserts) and only where local communities consent; Sacks remains skeptical such communities will volunteer.
- 58:00 – 1:11:10
Alternative Paths: Solar, Materials, and Grid Reform vs Nuclear Risk
Amid the nuclear clash, Chamath presents an alternative roadmap: improved material science, better storage, and grid reform could cut power costs without betting heavily on SMRs. Friedberg, however, underscores that countries with cheaper abundant electricity will win the AI and industrial race, and nuclear may be necessary to stay competitive.
- •Chamath argues we’re nowhere near AI being energy‑constrained; current bottlenecks are innovation and talent, not kilowatt‑hours.
- •He believes advances in materials, specialty chemicals, storage, and solar manufacturing can materially lower costs without nuclear’s regulatory and political baggage.
- •He emphasizes that the 20¢/kWh figure in some regions is more about corruption, legacy infra, and utility structure than generation technology.
- •Friedberg counters that if China pushes nuclear and drives down power to ~5¢/kWh while the US stays high, the competitive gap will be severe.
- •They touch on international SMR rollouts in China, India, and Russia as real‑world pilots that could eventually pressure the US to follow.
- •Jason suggests siting reactors 50–100 miles from population centers; Friedberg says transmission is manageable with adequate grid investment.
- 1:11:10 – 1:16:00
Lawfare, the California Coastal Commission, and SpaceX
The final segment spotlights the California Coastal Commission’s decision to restrict additional launches from Vandenberg, allegedly citing Elon’s political tweets as justification. Friedberg and Sacks use the case to illustrate how regulatory bodies can drift from their original mandates into politically motivated obstruction of critical infrastructure.
- •Coastal Commission originally created in 1976 to protect beach access and regulate coastal development.
- •Over time, its authority expanded to include veto power over activities like launches from Vandenberg Air Force Base.
- •A recent 6–4 decision to restrict more launches reportedly referenced Elon’s political tweets, which Friedberg calls ‘abhorrent’ and a distortion of the Act’s intent.
- •Sacks frames this as open ‘lawfare’—admin agencies punishing political opponents using regulatory discretion.
- •They note Elon has sued; Jason points out the Biden–Harris administration could theoretically rein in such behavior but hasn’t.
- •Brief return to 2024 stakes: Sacks warns lawfare escalates if not checked regardless of who wins; Jason notes Trump has also talked about using state power against opponents.
- 1:16:00 – 1:18:43
Wrap‑Up: Nuclear, Election Night, and Community Meetups
The show ends with more joking about nuclear site visits, Sax’s resistance to ‘progress,’ and election‑night scenarios if Trump loses. They plug global meetups for the 200th episode and tease the suspense (or lack thereof) around how each host will vote.
- •Jason jokingly proposes visiting a nuclear power plant; Sacks refuses, seeing no upside.
- •They rib each other about being pro‑ or anti‑progress and who’s truly ‘virtue signaling.’
- •Speculation about post‑election lawfare and whether Sacks would face reprisals in a Kamala administration.
- •Jason teases revealing his vote on the election special; others joke it will hardly be surprising.
- •Promotion of fan‑organized All‑In meetups around the world to celebrate 200 episodes.
- •Sign‑off with the usual affection and running jokes among the ‘besties.’
