All-In PodcastElon gets paid, Apple's AI pop, OpenAI revenue rip, Macro debate & Inside Trump Fundraiser
CHAPTERS
- 0:00 – 9:00
High-Stakes Cold Open: Live Blackjack with Tim Naki
The episode opens with Kiwi blackjack creator Tim Naki joining live from Calgary to play a single $10K blackjack hand on behalf of the All-In hosts and his fans. They win with a 21, turning $10K into $25K and decide to gift $15K to Tim and his fiancée to fly first-class to the All-In Summit.
- •Tim Naki’s backstory: Kiwi dairy farmer from Taranaki who moved to Calgary post-COVID and built a 1.3M-follower Instagram audience with one-hand-a-day blackjack videos.
- •All-In puts $10K into Tim’s daily hand as a “free roll” for Tim and his fans; Tim adds it to his own $14K bet for a $24K hand.
- •They hit a natural 21 against a live online dealer, turning the All-In $10K into $25K and Tim’s total win into $15K extra for him.
- •Hosts insist Tim keep the $15K to take his fiancée first-class to the All-In Summit, hinting at a casino night and more blackjack content.
- •Conversation about live-dealer gambling warehouses, influencer cross-promotion, and the degeneracy of wishing they’d bet more.
- 9:00 – 23:16
Influencer Dynamics and How the Collab Happened
After Tim drops off, the besties debrief how the surprise segment came together and reflect on the weird new world of micro-celebrities and mutual recognition across platforms. They joke about their own gambling impulses and the feeling of ‘losing’ by betting too little.
- •Tim is a genuine fan of the pod; a listener connected him to JCal who then invited him to do a live hand.
- •Modern influencers often know of each other “by default,” creating easy pathways for crossovers.
- •The hosts admit they initially thought Tim’s appearance might be faked or AI.
- •Discussion of gambler psychology: regret after winning a small bet instead of a bigger one.
- •Self-aware banter about being “degenerate” for feeling like they “lost 60K” by only staking $10K.
- 23:16 – 27:00
Inside Sacks’ Trump Fundraiser: Security, Spectacle, and Donor Mix
Sacks walks through the logistics and atmosphere of hosting a Trump fundraiser at his San Francisco home, from Secret Service sweeps to unexpectedly large pro-Trump crowds. He notes that many attendees were first-time Republican donors who previously supported Democrats.
- •Secret Service and police shut down multiple blocks; detailed advance planning included mapping Sacks’ house.
- •Media tried to gin up anti-Trump protests, but in reality there were far more pro-Trump supporters.
- •Intercom founder Owen McCabe observed that many donors did not identify as Republicans and had historically backed Democrats.
- •Key motivating issues: war, immigration, crypto, and innovation policy rather than party loyalty.
- •Campaign claims the event had more first-time donors than usual, signaling a shift among tech/finance elites.
- 27:00 – 35:30
Trump in the Room: Charisma, Policy Pivots, and Kamala Scenario
Chamath and Sacks describe Trump as unexpectedly charming, sharp, and funny, emphasizing how different he feels from media portrayals. They highlight his pro-innovation stances on AI and crypto, his energy level, and discuss Biden’s perceived decline and the possibility that the real contest is “Trump vs. Kamala.”
- •Trump spoke extemporaneously for about an hour with only a few topic notes, no teleprompter.
- •Anecdote: Trump jokes with the Winklevoss twins about “creating Facebook” and being “dealt a lot of cards,” impressing Chamath with his quick read of the room.
- •Chamath says he misjudged Trump in the past; found him polite, kind, and deeply engaging one-on-one.
- •Policy themes: pro-AI, pro-crypto, low regulation, low taxes, pro-America, contrasting sharply with Biden.
- •JCal argues Democrats are ignoring voter sentiment and fielding a candidate in obvious cognitive decline, predicting a Biden “demolition” unless there’s a “hot swap.”
- •Chamath posits the race is effectively Trump vs. Kamala Harris, raising fairness concerns if voters get a de facto successor they never vetted in a primary.
- 35:30 – 40:12
Biden, Enthusiasm Gap, and 2024 Electoral Calculus
The group debates polling, enthusiasm, and viability for Biden vs. Trump, including concerns over Biden’s health and the lack of excitement for his candidacy. They contrast intense visible enthusiasm for Trump with what they see as tepid or purely anti-Trump support for Biden.
- •They reference viral videos of Biden appearing frozen or disoriented and compare them to Mitch McConnell’s freezing episodes.
- •Discussion of whether calling out Biden’s condition is “elder abuse” vs honest assessment for the presidency.
- •Sacks questions what job Biden is now qualified for, arguing he wouldn’t be hired for any serious role outside politics.
- •They note FiveThirtyEight’s forecast showing a near coin-flip race despite Biden’s issues.
- •Consensus that Trump generates real excitement (e.g., UFC entrances, influencer support like Logan Paul), while Biden’s support is mostly driven by dislike of Trump.
- •Chamath and Sacks converge: enthusiasm is overwhelmingly on Trump’s side, which matters more than static polls.
- 40:12 – 45:00
Tesla Shareholders Reaffirm Elon’s $56B Package and Exit Delaware
Attention shifts to Tesla’s shareholder vote re-approving Elon Musk’s voided $56B pay package and the decision to re-domicile from Delaware to Texas. The hosts see the vote as a clear rebuke to judicial activism and a signal that shareholders value Musk’s continued leadership more than dilution concerns.
- •The comp package was initially considered nearly impossible to hit, which is why 73% of shareholders approved it in 2018 (excluding Elon and Kimbal).
- •A Delaware judge later voided the package, claiming shareholders weren’t properly informed; Tesla responded by putting it back to a vote.
- •Shareholders again approved with ~73%, identical to 2018, which Sacks sees as proof shareholders want to honor the deal.
- •Tesla’s stock popped on the news, suggesting markets value securing Elon’s presence more than the 10% dilution.
- •The move from Delaware to Texas is framed as part of a broader jurisdictional shift, alongside talks of a Texas Stock Exchange.
- 45:00 – 50:17
Executive Comp, ‘Scumbag’ Flip-Floppers, and Delaware’s Heist Problem
The panel criticizes institutional investors like CalPERS who voted yes in 2018 and no in 2024, arguing they effectively tried to back out of a deal after Elon delivered. They also condemn the lawsuit that voided Elon’s package as a trial-lawyer heist and warn that Delaware’s willingness to enable such actions threatens its corporate-law primacy.
- •Friedberg contrasts Musk’s performance-based package with “maintenance-mode” CEOs who earn tens of millions regardless of value creation.
- •He argues more boards should adopt Tesla-like pay structures to attract ambitious, risk-taking leaders.
- •JCal and Chamath call yes-then-no voters “morally and ethically void” and “scumbags,” suggesting founders/boards should avoid doing business with such institutions.
- •Clip of a CalPERS representative dodging a basic ethics question (“what would you do if your pay was pulled retroactively?”) illustrates perceived bad faith.
- •Sacks describes the lawsuit as a heist: trial lawyers recruited a tiny shareholder, got a judge to void the package, and are now asking for billions in fees—possibly paid in Tesla stock.
- •If Delaware awards massive fees for undermining a shareholder-approved package (now twice approved), they argue companies will flee to Nevada, Texas, etc.
- 50:17 – 55:45
Apple Intelligence: LLM Siri, App Agents, and a Privacy Trade
They analyze Apple’s WWDC reveal of “Apple Intelligence,” which will power new writing tools, call summaries, prioritized notifications, and a much smarter Siri that can act inside apps. While Wall Street rewarded Apple with a $300B+ market-cap bump, the besties question the substance and implications of Apple’s deep integration of OpenAI.
- •New features: Grammarly-like writing aids, AI call transcription and summaries, smart email/iMessage triage, and AI agents that execute tasks across apps (e.g., food orders).
- •Requires newer hardware (M1 Macs, iPhone 15+), which could drive an upgrade supercycle and re-anchor value in tightly coupled hardware+AI.
- •Chamath notes the shift from fully-baked devices to forward-looking, third-party software demos—a more “vaporware” Apple than the Jobs era.
- •Apple’s deal with OpenAI is reportedly non-cash and non-exclusive; Apple is also negotiating with Google, likely to auction off LLM defaults like search defaults today.
- •Sacks warns Apple is letting an external actor into the OS layer, contradicting its usual walled-garden/principled-privacy posture.
- •Apple claims strong consent flows (like location/photo prompts), but public trust in OpenAI lags trust in Apple, creating reputational risk.
- 55:45 – 1:02:00
OpenAI’s $3.4B Run Rate: Great Product, Questionable Moat
The group evaluates reports that OpenAI has hit a $3.4B annualized revenue run rate, driven by a mix of $20/month subscriptions and API usage. They praise ChatGPT-4o’s product quality, but question how much of the revenue is durable B2B vs fickle B2C, and how long proprietary edge can last against rapidly improving open-source models.
- •Sacks: OpenAI’s products are “awesome,” particularly GPT-4o and Whisper, with tangible performance gains over prior models.
- •He splits the business into three buckets: consumer subs, business subs (same product, different payer), and metered API for developers.
- •B2C SaaS historically suffers 5–10% monthly churn; real enterprise value lies in B2B/API with net dollar retention >100%.
- •Chamath doubts many enterprises are running mission-critical, production AI flows yet; most usage is experimental or “toy apps.”
- •Google Trends suggest US ChatGPT interest has dropped ~25% since end of school year, indicating heavy student usage, while international queries still trend upward but with lower ARPU.
- •Friedberg and others predict enterprises will increasingly prefer open-source LLMs (e.g., LLaMA) tuned internally, reducing dependence on expensive proprietary APIs.
- 1:02:00 – 1:05:26
Real-World AI Use: Productivity Boosts and Error Tolerance
Friedberg shares a concrete example of using ChatGPT across a 20-person offsite, where the team leaned on LLMs to gather data, create analyses, and accelerate prep work. They discuss cross-checking outputs across multiple models, hallucination rates, and the future shift from buying SaaS to building internal AI tools.
- •For a major offsite, Kohala used ChatGPT extensively for market sizing, data gathering, and unit-economics-style analyses, saving “hundreds of hours.”
- •Friedberg estimates roughly 25% of outputs were wrong or needed clarification, but this was acceptable for the strategic/approximate use case.
- •He validates the need to always pull sources and cross-check facts, often querying multiple LLMs (ChatGPT, Claude, Gemini) to triangulate truth.
- •The application layer is still mostly stand-alone chat; the next phase is embedding AI into internal data and workflows to replace many third-party SaaS tools.
- •JCal’s team similarly uses multiple LLMs to compute TAMs and other diligence inputs, normalizing for hallucinations by cross-referencing models.
- •Consensus: AI already makes knowledge workers “bionic,” but robust, internal, production-grade AI systems are still early in the enterprise adoption curve.
- 1:05:26 – 1:09:30
Macro Act I: Are We Actually Beating Inflation?
The hosts dissect recent economic data: CPI at ~3.3%, rapid prior rate hikes, 272K new jobs, and 4% wage growth. While JCal initially frames it as evidence that inflation is ‘broken’ and the landing is soft, Friedberg pushes back hard, pointing to low real GDP growth vs inflation and high borrowing costs as signs of stagflation-like stress.
- •Context: CPI fell from ~9% peak to ~3.3%, while the Fed hiked faster than any time in recent history.
- •New jobs beat expectations, and average hourly earnings are up 4% YoY vs ~3.x% inflation, suggesting a small real wage gain on paper.
- •Friedberg’s three key numbers: inflation, GDP growth (~1.3% in Q1 2024), and cost to borrow (~4.7%).
- •His argument: if the economy grows 1.3% while prices rise ~3%+ and debt costs ~4.7%, everyone’s real ability to pay (households, companies, government) deteriorates.
- •Elizabeth Warren’s public letter to Powell urges rate cuts, citing consumer and small business pain—evidence of political pressure on the Fed.
- 1:09:30 – 1:14:00
Macro Act II: Savings Burnoff, Job Openings, and Rate-Cut Politics
Chamath and Sacks argue that the drawdown of excess COVID savings and declining job openings signal looming economic softness. They see elevated unemployment risk, continued stagflation pressure, and a Fed chair more worried about his place in history than short-term political demands.
- •Chamath: consumers have largely burned through excess savings; now they must reenter the workforce, but companies are in defensive mode.
- •Unemployment is beginning to tick up; states like California show especially weak labor conditions.
- •Job openings have fallen from ~12M to far lower levels, suggesting a tighter job market with fewer opportunities for returning workers.
- •Sacks notes the gap between early-2024 expectations of 7 rate cuts and current market pricing of just 1, with Powell refusing to commit to timing.
- •Vinny Lingham’s point: a token 25 bps cut isn’t the issue—Powell doesn’t want to signal the *start* of a cutting cycle prematurely.
- •The panel agrees Powell’s primary incentive now is legacy: he wants to be seen more like Volcker (inflation slayer) than Burns (1970s inflation enabler), making him biased toward caution on cuts.
- 1:14:00 – 1:23:30
Macro Act III: Deficits, Election-Year Jobs, and Post-Election Risks
The discussion zooms out to how much current GDP depends on government deficit spending and election-year hiring. They argue that stripping out the 6% of GDP deficit would reveal recessionary conditions and question what happens to growth, jobs, and markets after the election once fiscal stimulus and political hiring normalize.
- •Friedberg: with a ~1.3% GDP growth rate and ~6% of GDP deficit spending, a balanced budget would likely mean negative GDP growth.
- •Sacks adds that many of the new jobs are in government, not the private sector, and that public hiring often spikes in election years.
- •The record Dow/Nasdaq levels are heavily concentrated in a few mega-cap AI names (NVIDIA, Microsoft, Apple), masking broader market softness.
- •JCal suggests Powell may have “stuck the landing,” but Chamath/Sacks think that’s premature with real economy stress building underneath.
- •Consensus that a negative GDP print going into the election is plausible, raising stakes for both fiscal and monetary policy choices.
- 1:23:30 – 1:24:47
Call for Help: Rare Sarcoma Case Search at Dana-Farber
The episode closes on a serious note as they amplify VC Peter Fenton’s urgent search for patients with a very rare sarcoma subtype to support a Dana-Farber study. They frame the effort as potentially practice-changing for future cancer treatment and urge any affected listeners or clinicians to come forward.
- •Condition: myxoid leiomyosarcoma featuring a PLAG1 fusion, an ultra-rare sarcoma variant.
- •Lead researcher: William Gibson at Dana-Farber, looking for genetically similar cases to build a meaningful cohort.
- •Personal stakes: the patient is Peter Fenton’s sister and the wife of their friend Pete Briger.
- •The study could “change the future of cancer” if enough comparable cases are found and analyzed.
- •They ask listeners or anyone in their networks with such a diagnosis to contact Fenton or the pod so they can connect them to Dana-Farber.
