Hurricane fallout, AlphaFold, Google breakup, Trump surge, VC giveback, TikTok survey

Friedberg explains that record-high Atlantic sea surface temperatures provide huge energy reserves that drive rapid hurricane intensification (e.g., Helene jumping from Category 2 to 4/5 in ~48 hours). A Nature paper shows that removing sulfur dioxide from shipping fuel—done to reduce acid rain—eliminated reflective ship tracks (clouds), allowing more solar energy into the oceans and likely doubling the rate of ocean warming in the 2020s. This ‘artificial cooling’ is gone, contributing to more frequent “1‑in‑100” or “1‑in‑500-year” storms now appearing every few years.

With $500B–$1T of real estate along Florida’s coast and ~$454B in mortgages, hurricane loss frequency assumptions have shifted from ~1-in-100/500 years toward 1-in-20–50. Insurance premiums required to fairly price that risk would reach 2–5% of a home’s value per year, which most homeowners can’t afford. Insurers are exiting, the state-backed reinsurance fund (Florida Hurricane Catastrophe Fund) has only $17B in statutory capacity versus events projected at $40–$50B+, and federal bailouts would set a precedent other states would demand. The besties argue many coastal markets (e.g., West Palm Beach, Malibu) are “massively overpriced” relative to this embedded risk.

DeepMind’s Demis Hassabis and John Jumper win the Nobel in Chemistry for AI-based protein folding. Friedberg explains that for decades, biologists struggled to predict 3D protein structures from amino acid sequences (e.g., Folding@home). AlphaFold essentially solves that prediction problem at scale and DeepMind has published structures for ~200M proteins. This enables rational design of new proteins for drugs, industrial enzymes, and nanoscale machines. It has already spawned a new Google venture (Isomorphic Labs) and dozens of startups, with billions in funding, though real-world impacts will unfold over several years.

Friedberg argues that large, cash-rich monopolies (Bell Labs/AT&T then; Google and Amazon now) have historically funded breakthrough R&D in ways smaller startups cannot—e.g., DeepMind/AlphaFold, Waymo, AWS. Chamath counters that elongated innovation cycles have weakened natural ‘creative destruction,’ letting giants sprawl into many markets and become irresistible political targets. The DOJ is now signaling structural remedies against Google, potentially forcing divestitures or separation of Chrome, Android, Play, and Search. The group suggests that while breakups may help ad-tech competition, they risk undermining long-horizon research capacity.

CRV’s decision to return or not call ~$275M of a $500M growth/opportunity fund is framed as rational: late-stage valuations remain high, exits are scarce, and the math of big growth funds doesn’t work. Sacks notes that with a $1B venture fund and ~10% ownership in a winner, you need a ~$30B outcome just to 3x the fund—an extremely rare event. LPs have over-“smeared” capital across too many GPs chasing AUM. Both Chamath and Friedberg say LPs are now preferring fewer managers and smaller funds, with growth capital more attractive now that crossover tourists (e.g., Tigers, hedge funds) have largely withdrawn.