E110: 2023 Bestie Predictions!

Late-stage startup valuations had already collapsed far more than headline marks admitted. Chamath said seven companies valued at $3–12 billion could not place convertibles because market-clearing prices were 80–90% lower, while structured warrants could preserve paper marks and eventually wipe out common shareholders and employees.

OpenAI was already being priced like a major platform, but its defensibility remained unsettled. As a reported $29 billion tender involving Founders Fund and Thrive surfaced, the hosts argued that models trained on the same public data may converge, making proprietary data crucial while commercial use of copyrighted training material could trigger lawsuits.

Google Search was identified as the incumbent most exposed to generative AI even if no single challenger wins. Chamath estimated that losing 10–15% of usage would materially hurt Google because several well-funded rivals can crawl the web and offer a simpler proposition: answers instead of search results.

Relativity Space’s January launch was framed as a binary company-making event. Chamath, an early investor, said success would unlock roughly a $10 billion order book and claimed full 3D printing could cut rocket build costs to $5–50 million, versus roughly $100–500 million for SpaceX.

Commercial real estate looked poised for a financing reckoning, not merely a remote-work slump. The discussion cited roughly 27% San Francisco office vacancy and Blackstone BREIT’s redemption restrictions; Chamath said UC supplied about $4 billion after Blackstone guaranteed 11.5% and posted $1 billion of its own equity as support.

Consumer credit was presented as the likeliest place where household strain would finally break. The hosts cited 19.6% average credit-card rates, mortgages above 7%, falling savings and still-heavy spending, with Friedberg expecting defaults to spread across cards, buy-now-pay-later products and mortgages.

Cell and gene therapy had become an infrastructure story as much as a medical one. Friedberg cited 27 FDA-approved therapies and more than 1,000 clinical trials, but treatments costing over $1 million because editing, handling and reinfusing cells requires complex bespoke systems that still take days or weeks.

Starlink’s addressable market was argued to be much larger than simply replacing home broadband. Chamath pointed to roughly one million subscribers and aircraft connectivity costing around $500,000 annually via Ka-band versus about one-tenth as much for Starlink, arguing the product also serves second connections and places without prior broadband.