E126: Big Tech blow-out, Powell’s recession warning, lab-grown meat, RFK Jr shakes up race & more

Google’s quarter was financially solid but strategically unresolved: cloud turned profitable for the first time, net income was $15 billion, and it authorized a $70 billion buyback. The sharper concern was that management still had not clearly explained its AI response or cost structure, despite anecdotes of faster internal releases.

The group’s broader thesis was that mega-cap tech had become cash-generating, low-growth businesses rather than primary sources of new platform innovation. Microsoft, Google and Meta were running at roughly 7%, 3% and low-single-digit revenue growth, while layoffs and buybacks increasingly supported earnings and share prices.

Jason was already using GPT-4 plugins and Zapier for LP research, social-profile discovery, outreach workflows and newsletter production, not just experimentation. He estimated roughly 30% of current knowledge work was automatable and had put employees at both companies on GPT-4 and the Playground.

The strongest AI argument was product creation, not job cutting: one self-taught coder reportedly assembled a voice-and-vision assistant in hours by combining existing tools. Friedberg’s point was that cheap, composable intelligence lets individuals build products that recently would have required a startup team.

San Francisco office values had reset brutally: 350 California Street, valued around $300 million four years earlier, was expected to attract bids near $60 million. With vacancy above 30% and potentially nearing 40%, the speakers thought buyers would need equity, patience and a five-to-ten-year horizon.

The same cheap-money reset was visible in venture capital: Chamath said 70% of companies that went public since 2020 traded below total cash invested since founding. He argued many private startups were therefore likely worth less than their preference stacks, leaving viable businesses with capital structures that no longer matched underlying value.

Recombinant animal proteins look technically workable but remain bottlenecked by economics and manufacturing capacity rather than basic biology. Friedberg said only about 1 million of roughly 100 million liters of global biomanufacturing capacity was rentable, while about 200 synthetic-biology startups competed for it, pushing capacity costs up roughly fourfold.

Cultivated meat is much further from mass-market economics because whole animal cells require suspension growth, scaffolding substitutes and costly signaling inputs. Costs remained orders of magnitude too high and may require years and billions more, while higher-value proteins such as rennet and pepsin were already being displaced by recombinant production.