What this covers
Michael Blastland and Russ Roberts explore the limits of predictability in science, medicine, and policy. Blastland's argument centers on The Hidden Half—a vast domain of variation that lies beyond genes and environment. He contends that even genetically identical organisms raised in identical conditions diverge dramatically, and that knowledge robust at the population level collapses when applied to individuals. Drugs proven effective in randomized trials help only a fraction of those who take them; smoking predicts lung cancer across countries but not within them; development interventions succeed in one place and fail in another. The conversation weaves through medicine, economics, psychology, and history to build a case that uncertainty is not a temporary gap in human knowledge but a permanent feature of how the world operates.
Blastland identifies the Weakest Link Principle as a central mechanism: a single overlooked factor—a mother-in-law's control over food, a photocopier requirement, an adoption threshold—can cause complete failure of an otherwise sound intervention. He illustrates this with examples from nutrition programs in South Asia, water supply schemes, and deworming campaigns. He also challenges common reasoning errors, particularly Survivorship Bias in entrepreneurship advice and Post Hoc Fallacy in medicine. Throughout, he argues for Epistemic Humility—acknowledging that overconfidence has preceded major shocks from Brexit to the 2008 financial crisis. Rather than viewing irreducible uncertainty as a problem to solve, Blastland suggests it is something to embrace, since Embracing Uncertainty is what gives life texture and allows competing visions to coexist in business and discovery.
Blastland argues that beyond nature and nurture there is a vast 'hidden half' of irregularity, chance, and irreducible variation that disrupts our ability to predict and control outcomes, and that intellectual humility about this uncertainty is both more accurate and ultimately something to embrace rather than fear.
- Genetically identical organisms in identical environments still vary enormously, proving a third source of variation beyond genes and environment
- Knowledge that is robust at the population level can be nearly useless at the individual level (drugs, smoking, aid programs)
- The 'weakest link' means robust interventions fail to travel across contexts because of unpredictable local details
Interventions fail to generalize because the limiting factor differs unpredictably across different implementation contexts.
- The reason interventions like deworming or education programs fail to generalize and scale is that the weakest link is not the same weakest link every time—it may be the mother-in-law, the photocopier, the soil, or the inability to leave the village for a job—making it nearly impossible to anticipate which factor will break the next implementation.
“the weakest link isn't the same weakest link every time, which makes it hard”
Top-down interventions ignore local knowledge and norms, imposing uniform solutions that cause unintended economic harm.
- Top-down development interventions are ethically and intellectually problematic because outsiders not only are imperfectly informed about local norms (e.g., why people grow yams not maize) but fail to respect them, and prescribing uniform solutions like everyone growing maize can cause gluts, collapsing prices, and starvation.
“We don't respect them. We say, 'Oh, that's the wrong way to fix it. Don't do that. This is our way, and our way is better.'”
People wrongly credit treatments for recoveries that would have happened anyway due to post hoc reasoning.
- Because many conditions resolve on their own, people wrongly attribute recovery to a treatment or intervention—a post hoc ergo propter hoc fallacy—as illustrated by a skin tag that disappeared on its own over two years that would have been credited to surgery had it been removed.
“many procedures, of course, are like that. People attribute any post hoc ergo propter hoc--after this, therefore because of this--classic fallacy”
Established assumptions enable progress but become liabilities when the assumed regularity sometimes fails in practice.
- Knowledge and effectiveness advance when we stop thinking about things we have established as reliably true (per Whitehead), which is a feature most of the time—freeing attention—but a bug when the assumed regularity is sometimes false.
“it was Alfred North Whitehead who said, actually, knowledge and effectiveness advances when we don't think about things.”