All-In PodcastBiden chaos, Soft landing secured? AI sentiment turns bearish, French elections
CHAPTERS
- 0:00 – 13:00
Cold Open, Banter, and Setup
The hosts open with lighthearted banter about Fox News appearances, Italian vacations, wine at the gym, and joking Star Wars references. They re‑establish their dynamic, briefly preview Sacks’ upcoming RNC speech, and introduce the episode’s docket, signalling an in‑depth focus on macroeconomics and AI.
- •Jason jokes about becoming a Fox News contributor and Sacks’ media presence.
- •Chamath calls in from Italy, riffing on summer, white wine, and exercise habits.
- •Sacks confirms he’ll speak at the RNC but withholds details about his theme.
- •Friedberg notes he’s been outdoors and sets a casual, conversational tone.
- •The group frames the main topics: macro conditions, AI, and global politics.
- 13:00 – 25:00
Has the Fed Achieved a Soft Landing?
The discussion turns to U.S. macro data: cooling inflation, rising unemployment off historic lows, and soaring equity indices driven by a handful of mega‑cap tech names. The hosts debate whether this constitutes a successful soft landing or masks looming recessionary pressures and unresolved structural issues.
- •June CPI prints at 3.0% YoY, below expectations; first MoM decline since 2020.
- •Unemployment edged up to 4.1% from a 3.4% low, signaling cooling labor markets.
- •Prediction markets assign a high probability (~90%) to rate cuts by September.
- •Chamath worries inflation may be falling because people are out of money, not purely due to Fed skill.
- •Friedberg notes P/E expansion is heavily driven by top tech names, not the broad market.
- •Powell is under pressure to consider 3% as a new normal instead of the historical 2% target.
- 25:00 – 38:20
Inflation Targets, Global Inputs, and Partisan Economic Perception
Friedberg challenges the realism of a 2% inflation target in a globalized, post‑COVID economy, arguing global input costs and stimulus overhang will likely keep inflation closer to 3%. Jason introduces research on partisan economic perception, showing how views of economic health increasingly track party affiliation rather than objective conditions.
- •Friedberg explains how global labor, commodities, and supply chains transmit foreign inflation into U.S. prices.
- •Food and CPG companies have taken double‑digit price hikes, and grocery prices are up significantly since COVID.
- •DC chatter includes redefining the Fed’s target from 2% to 3%, which Sacks argues will embed structurally higher rates.
- •Jason recounts how the 2% target originated somewhat arbitrarily in New Zealand and then became orthodoxy.
- •A chart on partisan economic perception shows Democrats and Republicans diverging sharply in their assessment of the economy depending on which party holds the presidency.
- •Friedberg ties perceived personal progress (income, housing, car purchases) to political narratives, suggesting perceptions can decouple from actual economic outcomes.
- 38:20 – 50:50
Market Concentration, AI Stocks, and Valuation Risk
Chamath introduces data showing extreme divergence between cap‑weighted and equal‑weighted S&P performance, matching extremes seen before the dot‑com crash. NVIDIA and a handful of AI‑driven giants dominate returns, prompting concerns about a late‑cycle hype peak and the need to reset risk‑return expectations.
- •Bloomberg data show the gap between cap‑weighted and equal‑weighted S&P is at its most extreme since March 2000.
- •NVIDIA and other “Magnificent Seven” stocks are responsible for much of the S&P’s gains.
- •Chamath warns we may be at the end of a major hype cycle similar to the dot‑com era.
- •Uncertainty about the true risk‑free rate and equity risk premium leaves investors “shrugging,” unsure how to price assets.
- •A reset could come through lower rates, lower equity valuations, or both, to realign incentives for productive investment.
- 50:50 – 1:05:50
Is the AI Boom a Bubble or Just Early Infrastructure?
The hosts interrogate the AI investment surge, contrasting skepticism about current ROI with optimism about long‑term transformative potential. Sequoia’s analysis and Goldman’s report suggest a large revenue shortfall versus AI capex, while the group debates whether this is a misallocation or a necessary overbuild akin to broadband in the 1990s.
- •Jason summarizes Sequoia’s argument that there’s a ~$500B “AI revenue hole” relative to projected capex.
- •Goldman’s report argues that near‑ to mid‑term AI productivity gains are limited by high costs and uncertain economics.
- •Chamath emphasizes AI’s impressive demos but points to hallucinations and narrow usable scopes; most current apps don’t justify trillion‑dollar spend.
- •He also flags NVIDIA hardware lock‑in as unsustainable, driving a rush toward alternative chips and architectures.
- •Friedberg attributes the bubble partly to ZIRP‑era capital still hunting for growth after SaaS and crypto cooled.
- •He notes that many AI metrics (model quality, token cost, energy per inference) are improving rapidly, suggesting future ROI could justify current infrastructure over a longer horizon.
- •Sacks counters “bubble doom” by invoking the internet and railroads: near‑term overbuilding that proved indispensable in the long run.
- 1:05:50 – 1:18:20
Real AI Use Cases, Labor Arbitrage, and Enterprise Adoption
Shifting from macro to concrete applications, the hosts discuss OpenAI’s revenue traction, future iPhone upgrades driven by on‑device LLMs, and early enterprise workflows. They see immediate labor‑arbitrage use cases, especially for research and knowledge‑work, even if generalized AGI‑like gains are distant.
- •Sacks cites OpenAI’s multibillion‑dollar run rate as proof of real market demand for LLM‑based “semantic search.”
- •He predicts a “snap upgrade” cycle when Apple ships an LLM‑powered Siri, assuming on‑device models protect privacy.
- •Jason recounts using GPT/Claude to aggregate salary data across sites, an example of replacing mid‑wage research tasks.
- •They foresee AI enabling one person to do the work of several, particularly in knowledge‑based roles, across startups and services like Athena.
- •Enterprise demand is currently strongest for “chat with your knowledge base” use cases, contingent on solving security and data‑access boundaries.
- 1:18:20 – 1:58:00
NVIDIA Dominance, GPU Clusters, and Andreessen’s Big Bet
The conversation narrows to GPU economics and venture strategy, using Andreessen Horowitz’s rumored 20,000‑GPU cluster as a case study. Chamath analyses the financial trade‑offs versus firm enterprise value, while Sacks and Jason consider whether such a cluster is smart BD, risky speculation, or just great PR.
- •Current high GPU prices are partly scarcity‑driven; Sacks expects prices to fall as supply increases.
- •Jason and Chamath describe how hyperscalers and NVIDIA have used GPU access as leverage in deal‑making with AI startups.
- •Andreessen’s 20,000‑GPU cluster could cost $400–600M; financed similarly to CoreWeave, the GP would still be on the hook for ~$100M+ in equity.
- •Chamath notes that spending $100M in management fees implies sacrificing potentially ~$2B in firm enterprise value (assuming ~20x multiples); any equity gained via the cluster must exceed that opportunity cost.
- •Sacks argues the cluster can be amortized by charging portfolio companies and serve as a business development and marketing asset.
- •Jason sees it as most valuable for very early startups that cannot yet buy/rent their own GPUs, and as a powerful PR move that has already paid off in attention.
- 1:58:00 – 2:15:00
Global Politics: French Elections, Populism, and Socialism Fears
Friedberg pivots to global politics, using the French election and recent votes in India, Mexico, and the UK to argue that socialist‑leaning policies are gaining momentum in reaction to inequality and globalization. Sacks and Chamath offer alternative explanations emphasizing anti‑incumbent waves, immigration backlash, and system‑level manipulations that thwart popular reform.
- •In France, a far‑left and centrist alliance blocked the surging right‑wing National Rally via tactical withdrawals in round two.
- •Friedberg outlines the left platform: 90% tax on income above €400K, lowering retirement age from 64 to 60, mandated 14% wage hikes, and price controls on essentials.
- •He sees similar signals in India (BJP losing majority), Mexico (Sheinbaum’s leftist win), and the UK—interpreting them as a broad push toward greater state involvement and redistribution.
- •Sacks frames French politics primarily as a revolt against unchecked immigration and cultural change, with elites gaming the system to deny nationalist reform.
- •He explains how Macron’s party and the far left cooperated to consolidate anti‑Le Pen votes district by district, giving a far‑left coalition the most seats despite National Rally winning the largest vote share.
- •Chamath suggests a simpler pattern: voters in many countries are throwing out incumbents after years of inflation, COVID disruptions, and economic pain, rather than coherently endorsing socialism.
- 2:15:00 – 2:35:00
Culture, Immigration, and Relative Deprivation
The hosts broaden from electoral mechanics to cultural and psychological underpinnings: resentment over inequality, social‑media‑driven comparison, and Western discomfort with enforcing assimilation norms. Chamath invokes Peter Thiel’s reading of the Ten Commandments to argue that constant lateral comparison fuels political anger and ideological swings.
- •Friedberg worries that youth in industrialized countries increasingly embrace socialist ideas as solutions to perceived rigged hierarchies.
- •Sacks argues voters primarily want reform and national sovereignty, not doctrinaire socialism, but elites prefer to ally with radicals rather than moderate on immigration.
- •Chamath observes that many non‑Western countries actively preserve traditional dress, language, and culture, while Western elites often denigrate their own traditions.
- •Jason recalls how the U.S. once explicitly taught a “melting pot” model of assimilation, which is now often criticized.
- •Chamath cites Thiel’s idea that most commandments warn against looking sideways and coveting others’ lives; social media amplifies this coveting and discontent.
- •They suggest that greater authenticity and less curated online performance might alleviate some of the perceived inequality fueling political radicalization.
- 2:35:00 – 3:05:00
Biden’s Cognitive Health, Media Complicity, and the ‘Speed Run Primary’
In a fiery close, the hosts allege that the White House and mainstream media have long concealed Biden’s cognitive decline. Jason calls for whistleblowers and revives his “speed run primary” concept, while Sacks catalogues what he calls a series of Democratic “hoaxes” against Trump and predicts the party will likely settle on Kamala Harris rather than an open contest.
- •Jason cites a neurologist’s White House visits and the press secretary’s evasive responses as signs of a cover‑up.
- •Sacks accuses journalists and party insiders of dereliction, noting that Clooney, Olivia Nuzzi, and Chuck Todd all admitted they’d known about Biden’s condition earlier but stayed quiet.
- •Dean Phillips is praised for publicly stating months ago that Biden was clearly in decline and should not run again.
- •Jason urges potential White House or medical whistleblowers to release Biden’s cognitive test results, arguing it’s unethical to hide them.
- •Jason’s “speed run primary” idea: Biden steps down in summer; a short, intense primary sprint with a handful of candidates, celebrity‑moderated forums, and ranked‑choice delegate voting before the convention.
- •A Semaphore article is presented as echoing this idea; Sacks doubts Democrats will risk the chaos and expects they’ll default to Kamala Harris if Biden steps aside.
- •Sacks frames the Biden issue as just one in a chain of Democratic “hoaxes” (e.g., Steele dossier, “very fine people” narrative), arguing Trump deserves another term as a rebuke.
- •They close with more joking around about predictions, Trump’s debate performance, and media strategy before heading into the show’s plug section.
- 3:05:00
Outro and Promotions
The episode ends with the usual playful sign‑off and an extended plug segment. They advertise the All‑In Summit, social channels, various host ventures, and partner companies, reinforcing the show’s ecosystem around startups, AI, and tech.
- •Hosts sign off with humorous banter and recurring in‑jokes about their dynamic.
- •They promote the All‑In Summit in Los Angeles and encourage applications.
- •Listeners are urged to subscribe on YouTube and follow the show on X, TikTok, Instagram, and LinkedIn.
- •Individual host projects are plugged: Chamath’s Substack, Sacks’ Glue.ai, Friedberg’s Ohalo Genetics, Jason’s Founder University, Launch fund, and Athena virtual assistants.
- •Jason asks listeners to share the podcast with two friends to grow the audience.
