Sequoia's David Cahn is backing Flapping Airplanes, a research lab founded by Ben and Asher Spector, ages 25 and 26, alongside Thiel Fellow and ex-Neuralink engineer Aidan Smith. The thesis is direct: data, not compute, is now the binding constraint on AI progress. Today's models are data-inefficient, a point Dwarkesh Patel has pressed in recent interviews with Andrej Karpathy and Richard Sutton. Flapping Airplanes is built around solving that specific problem, drawing on biological models of intelligence as its design principle.
The lab is a deliberate structural bet against the current industry consensus. Cahn argues that concentrating top talent on scaling work, rather than fundamental research, risks pushing AGI timelines out, not in. Flapping Airplanes is designed to replicate the independence and long time horizons of a Ph.D. program while closing the pay gap with Big Tech. The team recruits people who would previously have gone into quant finance or doctoral programs, not credentialed AI brand names.
The piece is worth reading in full not for its conclusion but for Cahn's framing of a core tension: scaling-first labs optimize for 1 to 2 year wins, while research-first labs must tolerate bets that take 5 to 10 years and carry low individual probability of success. Cahn's earlier writing flagged that corporate incentives structurally suppress the radical ideas scientific breakthroughs depend on. Flapping Airplanes is his answer to that structural failure.
[READ ORIGINAL →]