296 | Brandon Ogbunu on Fitness Seascapes and the Course of Evolution
What this covers
Sean Carroll speaks with Brandon Ogbunu, an evolutionary biologist at Yale, about the conceptual and mathematical tools needed to understand evolution as it actually occurs rather than as textbook simplifications suggest. Ogbunu's central move is to replace the Fitness Landscape metaphor—which posits that organisms climb toward stable peaks of reproductive success—with the Fitness Seascape, where the topography itself constantly shifts as environments change. This reframing exposes why classic evolutionary theory, built on equilibrium assumptions and deterministic prediction, has failed to solve contemporary problems: understanding drug resistance, engineering better crops, predicting which genetic variants cause disease. The conversation ranges across what blocks real prediction and what newer approaches might enable it, while also excavating the ethical dangers of acting as if we can predict and control evolution when we cannot.
The discussion opens onto several territories. Ogbunu explains how epistasis—interactions where two mutations produce unpredictable combined effects—generates a rugged, non-equilibrium landscape that simple gradient-descent models cannot navigate. He shows how subtle environmental shifts (even different antibiotics) redirect which mutations spread, making evolution context-dependent in ways traditional population genetics ignores. He separates what genomic data can and cannot do: rich association patterns emerge from sequencing, but association is not mechanistic understanding, and Correlation vs Causation remains a hard wall. Ogbunu identifies three systemic problems in modern biology—physics envy, uncritical categorization, and what he calls "giddiness" around data—that together produce overconfidence in embryo selection and other predictions about complex traits. He contrasts this with Frances Arnold's directed evolution and Complexity Science approaches that do offer genuine steering power, though not full control. Throughout, the conversation treats prediction and ethics as technically entangled: past disasters like eugenics were not only immoral but built on bad biology, and present overconfidence about engineering traits rests on the same kind of mechanistic ignorance.
Ogbunu argues that the classic fitness landscape metaphor must be upgraded to a dynamic 'fitness seascape' because real biology is non-equilibrium, epistatic, and environment-dependent, and that combining new genomic data with complexity-science thinking is bringing evolution closer to genuine prediction and control while exposing how overconfident, deterministic prediction is both technically wrong and ethically dangerous.
- Environments constantly fluctuate, so fitness landscapes are really shifting seascapes that change which mutations are beneficial
- Epistasis and weak genotype-phenotype mapping make evolution rugged and hard to predict with simple deterministic tools
- Post-genomic data plus complexity-science frameworks promise predictive and steerable evolution, but 'physics envy' and 'giddiness' make naive prediction (e.g. embryo selection, RFK-style denialism) misleading
Philosophy should anticipate and guide ethical questions rather than retrospectively judge decisions already made.
- Philosophy is supposed to be a source of wisdom about ethics, but it tends to be a trailing indicator—waiting for things to happen and then judging them bad—whereas with potentially scary genetic-engineering possibilities we need leading ways of thinking about things before they happen.
“philosophy is supposed to be a source of wisdom about ethics and morals and things, but I kind of think it's a trailing indicator.”
Incidental descriptive detail about the speaker.
- Sean Carroll's hair color is a dull red-brownish.
“It's kind of brownish. It's red brownish. Kind of dull.”