All-In PodcastE118: AI FOMO frenzy, macro update, Fox vs Dominion, US vs China & more with Brad Gerstner
CHAPTERS
- 0:00 – 6:35
Bestie banter from Japan: food, poker, and the quorum check
The episode opens with playful banter about travel in Japan, “austerity” dinners, and the show’s lineup with David Friedberg absent. Jason introduces guest Brad Gerstner as the “fifth Beatle,” setting up a packed agenda that will move from private markets to macro and geopolitics.
- •Jokes about “austerity 2023” and caviar ‘on the bottom’
- •Jason in Japan; talk about travel, exchange rates, and food culture
- •Friedberg is missing due to Wall Street commitments
- •Brad Gerstner joins as guest to discuss markets and macro
- 6:35 – 8:48
AI fundraising FOMO: platform shift vs. spray-and-pray investing
The group digs into the generative AI investment boom and whether it resembles prior platform shifts like the internet or mobile. They debate Doug Leone’s warning about FOMO, the inevitability of many failed experiments, and where long-term value capture will actually land.
- •Generative AI explosion: hundreds of startups and billions raised
- •AI framed as a major platform shift—but likely overestimated short term
- •Value capture uncertainty: foundation model winners vs. applications layer
- •Ecosystem dynamics: many experiments die; a few become category leaders
- 8:48 – 20:22
Where the money is: risk-free rates, compute ‘picks & shovels,’ and data moats
Chamath argues that higher risk-free yields raise the bar for venture returns, increasing pressure to deploy capital and potentially torching money in correlated bets. The conversation shifts to durable winners: custom silicon/compute, and unique proprietary datasets that can differentiate models and applications.
- •Higher cash yields force venture to target much higher long-term returns
- •OpenAI/Microsoft deal seen as a ‘tell’ about capped upside and competition
- •‘Gold rush’ analogy: compute/silicon providers (NVIDIA/AMD) as picks & shovels
- •‘White truffles’ thesis: scarce, high-quality datasets as enduring moats
- 20:22 – 25:22
OpenAI’s API price cuts and the coming wave of AI-enabled SaaS
Sacks explains why OpenAI lowering API costs can accelerate an explosion of AI features across existing software products. They discuss how companies can integrate LLMs without building models, and how many startups will rebrand as “AI companies” by adding one feature on top of APIs.
- •What an API is and why it matters for developers and distribution
- •Example integration: Notion using LLMs for tasks, tables, and summaries
- •OpenAI cutting API prices ~90% as a catalytic adoption move
- •Founders/VCs re-categorizing: ‘AI company’ as a marketing and fundraising label
- 25:22 – 29:06
LP reality check: allocations tighten and the power law reasserts itself
Jason asks what limited partners are thinking given attractive short-term yields. Brad argues LPs generally keep long horizons, but are narrowing commitments toward top-tier funds with proven DPI, leaving newer or subscale managers in a tougher fundraising environment.
- •LPs don’t rapidly reallocate like individuals; institutional pacing matters
- •Expectations: rates likely fall eventually, but near-term uncertainty persists
- •‘Narrowing the aperture’: capital concentrates into top-performing firms
- •Explosive fund formation in recent years sets up a shakeout among newer funds
- 29:06 – 37:35
Down rounds and board dynamics: cramdowns, pay-to-play, and founder priorities
The discussion turns to private-market cleanup: down rounds, structured financings, and why some VCs leave boards. They debate whether cramdowns can be done gracefully, and emphasize resetting valuations and refreshing employee equity to preserve team motivation.
- •Cramdowns/pay-to-play mechanics and why they’re becoming more common
- •Leaving a board is read as a serious signal vs. internal partner swaps
- •Chamath: founders must prioritize employees; reset valuation + refresh option pool
- •Brad: ‘tough talk’ and capital discipline were normal in earlier cycles
- 37:35 – 48:18
Macro whiplash: bond market signals, sticky inflation, and uncertainty risk
Sacks asks why the bond market shifted toward ‘higher for longer,’ prompting Brad and Chamath to weigh inflation’s stickiness and upcoming prints. Chamath highlights low-quality earnings and the cash-flow vs. reported income gap, while Brad stresses that uncertainty itself can freeze capital allocation.
- •Two-year yield swing interpreted as ‘Fed must keep hiking’
- •Brad: inflation is slowing but sticky; more hikes now expected
- •Chamath: record gap between reported earnings and cash from operations
- •Core risk: uncertainty shuts down investment more than any single rate level
- 48:18 – 52:01
Tech vs. the broader economy: withdrawal from ZIRP and where pain shows up
Jason notes seed-stage deal quality improving as discipline returns, while Sacks separates tech’s post-2021 reset from the wider economy’s lagging effects. They discuss the speed of rate changes, real estate impacts, and why innovation continues regardless of macro conditions.
- •Early-stage: better pricing, longer diligence cycles, milestone-based funding
- •Sacks: tech already crashed; broader economy still faces ‘withdrawal’ pain
- •Fast rate hikes transmit via housing, credit cards, and financing conditions
- •Brad: focus on durable growers rather than timing macro prints
- 52:01 – 1:02:05
Benioff’s ‘inner Elon’: efficiency drives, layoffs, and CEO agency
Salesforce’s shift toward profitability and buybacks leads into a broader conversation about Silicon Valley’s new ‘efficiency’ mindset. Brad and Sacks argue that Elon’s Twitter overhaul reminded CEOs of their agency to simplify orgs, reduce layers, and push accountability—though public companies have limits.
- •Salesforce pivots to profitability; buybacks expand amid activist pressure
- •Brad: Silicon Valley gained ‘courage momentum’ after Twitter/Musk actions
- •Sacks: CEOs often act like ‘prisoners’ of their companies; agency is underused
- •Practical limits: public companies can’t copy Twitter’s extremes without breaking execution
- 1:02:05 – 1:12:03
Stock-based compensation ‘grift’: adjusted EBITDA, dilution, and comp committee fixes
Chamath and Brad argue that SBC has been used to hide true labor costs and enable headcount bloat, especially when rates were near zero. Brad traces the GAAP vs. adjusted reporting history, criticizes comp consultants, and proposes tying incentives to free cash flow per share and strict dilution limits.
- •SBC treated as ‘non-cash’ in adjusted metrics enables profit overstatement
- •Employee counts and SBC exploded post-2015; dilution transferred shareholder upside
- •Examples: extreme SBC ratios and buybacks ‘round-tripping’ cash to offset issuance
- •Proposed remedy: comp tied to FCF/share; target low annual dilution (e.g., ~50 bps)
- 1:12:03 – 1:21:16
Fox vs. Dominion: defamation standards, ‘actual malice,’ and speech vs. lies
The group debates Fox’s exposure in the Dominion lawsuit and the broader implications for media accountability. Sacks argues the world would be better with greater liability for knowingly false claims, but current doctrine (NYT v. Sullivan) makes it hard unless actual malice is proven; he calls for reform.
- •Murdoch deposition: admits Fox could have stopped hosts but didn’t
- •Sacks differentiates Fox hosts who endorsed vs. skeptically challenged claims
- •Legal framework: NYT v. Sullivan and the high bar of ‘actual malice’
- •Policy view: reform libel/defamation law to penalize knowingly false reporting
- 1:21:16 – 1:34:32
TikTok and US–China decoupling: roadkill of great-power competition
Brad (a ByteDance shareholder) frames TikTok as a proxy battleground for a larger US–China decoupling and national security debate. Chamath predicts a ban despite preferring free markets, while Sacks emphasizes ‘great power competition’ as the organizing principle that will sweep TikTok into broader policy shifts.
- •Brad discloses ByteDance stake; argues TikTok is small vs. ByteDance economics
- •Debate: ban vs. forced sale/spin; US device bans already rolling out
- •Bigger theme: hard vs. soft decoupling; widening definition of ‘national security’ industries
- •GPC framing: chips, batteries, energy, supply chains become strategic arenas
- 1:34:32 – 1:46:36
Ukraine and China’s ‘peace proposal’: proxy-war framing and a heated clash
Sacks interprets China’s peace proposal as a PR move to claim moral high ground while the US dismisses negotiations because it is effectively a co-belligerent. The conversation escalates into a sharp argument about NATO expansion, provocation, and the costs of the war, with Jason defending a pro-democracy Western stance and Sacks pushing a realist strategy.
- •China peace proposal as diplomatic maneuver; US seen as rejecting mediation
- •Sacks: Ukraine war benefits China (US distraction, depleted stockpiles, discounted Russian energy)
- •Jason vs. Sacks dispute: democracy-spreading vs. realist security constraints and escalation risks
- •Debate over neutrality/Crimea concessions and who should decide (Ukraine vs. West)
- 1:46:36 – 1:56:24
Culture coda: Harvard legacy admissions, Draymond Green on Black History Month, and wrap-ups
The episode closes with two cultural topics: a viral claim about the share of Harvard students who are legacies/athletes/donor-related, and a clip of Draymond Green arguing for teaching Black history year-round rather than confining it to a month. Chamath also issues a quick correction to a prior Stripe chart before the group signs off with jokes and callbacks.
- •Harvard legacy admissions debate and perceived unfairness/meritocracy tension
- •Draymond Green clip: teach Black history from Jan 1 to Dec 31
- •Stripe chart errata: net vs. gross revenue line correction; valuation implications unchanged
- •Playful sign-off and post-argument ribbing among the hosts
