Patrick Collison — Why Silicon Valley's most talented should leave

Collison suggests that while SF is great for certain entrepreneurs, its culture over-valorizes startups and under-rewards long, quiet accumulation of expertise; many people would create more value and fulfillment by becoming world-class scientists or engineers instead of 23‑year‑old founders.

Across domains, people who worked under exceptional mentors or in very high-bar labs/teams report permanent upgrades in what they consider ‘great’; Collison advises 20‑somethings to deliberately seek places where excellence is embodied and can be absorbed tacitly.

Fast Grants surveys showed 79% of scientists would change their research ‘a lot’ if funding were flexible, implying huge misallocation under current grant structures; Collison argues we over-index on NIH budgets and under-index on micro-level constraints, evaluation norms, and career paths.

Arc Institute funds scientists, not projects; builds shared technical platforms; and offers non-PI scientific careers—this structure plausibly enabled risky work like “bridge editing” that might have been rejected by NIH, echoing DARPA’s and others’ role in CRISPR and mRNA.

Stripe’s existence in a supposedly moat-heavy, regulated domain shows that ‘defensibility’ is often less binding than incumbents believe; Collison sees cultural/organizational will and deep understanding of the domain as more decisive than theoretical network or regulatory moats.