All-In PodcastE173: Google buying HubSpot? FTX depositors not made whole, AI job fears, Ukraine joining NATO
At a glance
WHAT IT’S REALLY ABOUT
Google Eyes HubSpot, FTX Fallout, AI Jobs, NATO’s Ukraine Gamble
- This episode of the All-In Podcast introduces new All-In CEO Jon Hale, then dives into major stories spanning crypto, tech M&A, AI labor disruption, and geopolitics.
- The besties correct their earlier claim that FTX depositors were made whole, explaining how bankruptcy mechanics and timing left many crypto holders effectively short-changed despite legal ‘full repayment.’
- They debate the logic and antitrust risks of Google potentially acquiring HubSpot, the real economic impact of AI on white‑collar jobs and robotics, and the escalating stakes of Ukraine’s path toward NATO membership.
- Underlying themes include mistrust of mainstream media narratives, the constraints of U.S. debt and antitrust policy, and whether current leaders are sleepwalking into larger conflicts abroad.
IDEAS WORTH REMEMBERING
5 ideasFTX depositors are *not* truly made whole despite legal framing.
Bankruptcy law fixed customer claims in U.S. dollars at the November 11, 2022 bankruptcy date—when many tokens (e.g., Solana at ~$16) had crashed. The estate is now selling those same tokens into a massive run-up (Solana near ~$188) and using the proceeds to pay back claims at the old lower prices. Legally, this satisfies dollar-denominated claims; economically, depositors lost the entire upside and, in token terms, may receive a fraction (e.g., ~10 of 100 Solana). Remaining surplus is expected to flow to other creditors and ultimately equity holders, not to restore lost crypto upside.
Crypto exchanges sit awkwardly between ‘custodial wallet’ and ‘brokerage account’ models in bankruptcy.
Friedberg explains that traditional brokerages denominate accounts in a local currency and treat holdings as a portfolio to be liquidated into that currency at a point in time. Many crypto users, by contrast, believed they had *custodial* asset accounts, not trading accounts destined for fiat conversion. This mismatch led courts and trustees to apply fiat-based, Chapter 11-style liquidation rules that disregard users’ expectations of holding specific coins rather than their dollar equivalent at a crash low.
Media narratives around FTX and Russia–Trump links often carry agendas and require skepticism.
Sacks argues that mainstream coverage repeated the ‘FTX customers made whole’ line and may be seeding public support for a future SBF commutation or pardon, given his political donations. He also criticizes a Guardian story alleging Russian-linked funding to the Trump Media SPAC as evidence-light, noting the article itself admits there’s no indication Trump or the company knew the loan’s nature. He connects this to the Hamilton 68 ‘Russia bot’ project—which Twitter internal emails called “bullshit”—as an example of manufactured Russia scares feeding thousands of credulous media stories.
A Google–HubSpot deal would be strategically about advertiser lock‑in and data, but faces major antitrust friction.
Friedberg sees strong logic: Google’s $250B+ ad business generates leads; HubSpot manages and converts those leads via CRM and marketing automation. Deep integration could (1) improve conversion tracking and ROI, driving more ad spend, and (2) let Google use first-party CRM data to retarget and upsell, effectively becoming a ‘roach motel’ for advertisers. Chamath and Sacks counter that synergy may be overestimated, and that any large deal ($35–40B) would face intense FTC/EU scrutiny, especially around data use and market power, likely dragging on for years.
AI is likely to supercharge knowledge work productivity rather than cause immediate mass unemployment.
Chamath cites historical data showing that while specific job categories can be wiped out (e.g., farm to factory), overall GDP composition and aggregate worker compensation tend to remain resilient and track productivity gains over decades. Friedberg analogizes AI to the PC/software revolution: architects, engineers, and designers didn’t disappear when computers arrived; their output multiplied. With AI agents, a single knowledge worker could do 10–100x more (e.g., generating multiple architectural designs per day instead of one in weeks), lowering costs and expanding demand rather than simply eliminating roles.
WORDS WORTH SAVING
5 quotes‘According to the judicial proceedings, you’ve been “made whole,” but the truth is that Solana, at this moment, is trading at $188, so you have not been made whole, and this is why the crypto community is furious.’
— David Sacks
‘Bankruptcy rules are very specific and they’re completely designed around assessing what the value of each claimant is *at the time of bankruptcy*… It’s simply wrong or misleading to say that the depositors were made whole.’
— David Sacks
‘When productivity goes up, costs go down, the actual volume balloons and the economy grows… these systems are going to give humans 10 to 100x leverage, not simply replace knowledge work.’
— David Friedberg
‘It’s still more artificial than intelligent… everybody needs to take a deep breath and understand there’s just going to be a lot more work before you get to this omnipresent agent that just replaces and destroys everything and thinks on its own.’
— Chamath Palihapitiya (paraphrasing Yann LeCun)
‘If you want to have a serious chance of World War III in the next four years, then I would say go ahead and vote for Biden in November… I’m personally not willing to accept even a 1% risk of that.’
— David Sacks
High quality AI-generated summary created from speaker-labeled transcript.
Get more out of YouTube videos.
High quality summaries for YouTube videos. Accurate transcripts to search & find moments. Powered by ChatGPT & Claude AI.
Add to Chrome