Modern WisdomThe Problem With Trying To Be Rational - Steven Pinker
CHAPTERS
- 0:00 – 0:27
Reasoning has costs: acting with imperfect information
Pinker opens with the idea of bounded rationality: reasoning takes time and resources, so the “optimal” decision must be balanced against the costs of delay. You can’t gather data forever—inaction is itself a risk.
- •Reasoning requires time, information, and cognitive resources
- •Decision quality must be traded off against the cost of further analysis
- •At some point you must act under uncertainty
- •Inaction can be more costly than a suboptimal action
- 0:27 – 2:03
Why we need cognitive-bias tools in novel situations
Chris asks why modern life seems to require a toolkit of mental models and bias names to function. Pinker argues humans reason well in familiar, personally meaningful contexts, but struggle to generalize across abstract or new problems without explicit frameworks.
- •Humans are good at everyday causal/logical reasoning in familiar domains
- •Abstract, novel situations expose systematic errors (e.g., sunk cost, availability)
- •Naming fallacies and learning normative models helps transfer reasoning
- •General-purpose reasoning tools require conscious reminders
- 2:03 – 3:15
Can studying biases make you more rational (and what outcomes track rationality)?
Chris references Kahneman’s claim that learning about biases didn’t make him more rational. Pinker says self-assessment is hard due to the “bias blind spot,” but notes evidence that lower susceptibility to common fallacies correlates with better life outcomes.
- •Bias blind spot: we see bias in others more easily than in ourselves
- •Some improvement is possible, but hard to verify introspectively
- •More rational people tend to have better real-world outcomes on average
- •Reduced risk of scams, accidents, job/relationship problems
- 3:15 – 4:30
Intelligence vs. rationality: where smart people still fail
Pinker explains that intelligence and rationality are correlated but far from identical. Smart people can be especially vulnerable when beliefs tie to identity, tribe, or sacred values—fueling “my-side bias” and motivated reasoning.
- •IQ and rationality correlate imperfectly
- •Identity-linked beliefs resist updating with evidence
- •My-side bias: steering reasoning to protect the tribe/coalition
- •Keith Stanovich’s idea of an “irrationality quotient”
- 4:30 – 6:49
How to counter motivated reasoning: dissent, heterodox inputs, and free speech norms
Asked what smart people should watch for, Pinker details biased assimilation and “spin-doctoring” contrary evidence. He emphasizes social and institutional solutions—exposing ideas to criticism and seeking information outside one’s habitual sources.
- •Motivated reasoning patterns: cherry-picking and nitpicking counterevidence
- •Biased assimilation reinforces existing commitments
- •Seek out opposing or unfamiliar media sources intentionally
- •Communities with open criticism and free speech help correct individual bias
- 6:49 – 9:42
The rationality community’s methods: careful literature review as a model
Pinker cites writers within the rationality community—especially Scott Alexander—as examples of thorough, even-handed synthesis on politicized topics. The focus is on showing work, surveying evidence, and resisting pre-made conclusions.
- •Reputation for objectivity comes from transparent evidence synthesis
- •Literature reviews can de-politicize contentious questions
- •A good method starts with uncertainty rather than a fixed stance
- •Example discussed: assessing COVID lockdown effectiveness by reviewing studies
- 9:42 – 15:52
Bayesian reasoning in everyday life: priors, likelihoods, and base rates
Chris asks how to use Bayesian reasoning day-to-day. Pinker breaks Bayes’ rule into its practical components—priors, likelihood, and the overall prevalence of evidence—using medical testing and COVID examples to show why base-rate neglect leads people astray.
- •Beliefs should be calibrated on a continuum (0–1) rather than binary
- •Posterior belief updates from: prior × likelihood ÷ evidence commonness
- •Medical testing illustrates false positives and the importance of base rates
- •People reason better with natural frequencies than abstract probabilities
- 15:52 – 18:45
Forecasting and prediction markets: Bayes as a discipline of being “less wrong”
They connect Bayesian updating to forecasting tournaments and prediction markets. Pinker argues strong forecasters start with base rates and then adjust with specifics, while punditry often fails due to ideological distortion; markets help because accuracy is rewarded.
- •Superforecasting improves predictions via base rates and incremental updates
- •Prediction markets often outperform typical expert commentary
- •Pundits underperform when ideology overrides situation-specific evidence
- •Skin in the game incentivizes information gathering and honest calibration
- 18:45 – 21:30
Expected utility in real life: weighing risk, payoff, and uncertainty
Pinker shifts to rational choice and expected utility—multiplying probabilities by payoffs to guide decisions. He notes real life often involves uncertainty (unknown probabilities), but structured thinking still improves choices from driving risks to consumer purchases.
- •Expected utility: choose options with best probability × payoff totals
- •Distinction between risk (known odds) and uncertainty (unknown odds)
- •Everyday examples: speeding vs. arriving slightly earlier
- •Consumer example: extended warranties often favor sellers, not buyers
- 21:30 – 22:20
Insurance, self-insurance, and when protection actually makes sense
Continuing the utility discussion, Pinker explains why insurance is rational for catastrophic losses but often irrational for minor, replaceable items. The key is whether a loss is recoverable without financial ruin.
- •Insurance is most rational for rare, catastrophic outcomes
- •Extended warranties resemble “insurance for your toaster”
- •Self-insurance is better for manageable replacement costs
- •Companies profit because the average expected value favors them
- 22:20 – 24:29
Rationality vs. intuition: overthinking, ‘go with your gut,’ and bounded rationality
Chris describes getting stuck in his head when applying mental models. Pinker critiques the popular ‘gut instinct’ idea (à la Blink) as not generally reliable, while acknowledging real-life decisions often lack explicit probabilities and require judgment under constraints.
- •Overthinking is real, but intuition isn’t a universal shortcut to truth
- •Many life choices lack stated probabilities, forcing informed guesswork
- •Default advice: avoid impulsive action; think twice when possible
- •Bounded rationality: reasoning itself consumes scarce resources
- 24:29 – 32:19
Using other people’s outcomes as data: decisions, imagination limits, and cognitive dissonance
Pinker shares Daniel Gilbert’s advice: people are poor at forecasting future feelings, so it’s often better to look at how similar decisions worked out for others. He illustrates with his move from MIT to Harvard and discusses how cognitive dissonance can increase satisfaction post-choice.
- •Affective forecasting errors: imagination is an unreliable guide
- •Use real-world outcomes from similar people/situations as evidence
- •Pinker’s MIT-to-Harvard decision informed by colleagues’ experiences
- •Cognitive dissonance can reduce regret by rationalizing chosen paths
- 32:19 – 37:50
Conspiracy theories as unfalsifiable, moralized ‘memes’ that evade refutation
Chris asks how conspiracies subvert rationality. Pinker explains they’re often structured to be unfalsifiable (‘lack of evidence is proof’), spread like contagious memes, and function as moral/tribal signaling more than truth-seeking.
- •Unfalsifiability: counterevidence gets reinterpreted as part of the cover-up
- •Conspiracies spread because they’re cognitively ‘self-sealing’
- •Often serve my-side bias: identifying villains and affirming group identity
- •Many beliefs persist because they signal values, not because they’re evidence-based
- 37:50 – 41:02
Distrust in institutions, probabilistic plausibility, and how experts can rebuild credibility
They discuss how eroding institutional trust broadens what people will accept as ‘conspiracy.’ Pinker distinguishes real secrecy from outlandish multi-step plots via probabilistic reasoning, and argues institutions worsen distrust when they politicize themselves or fail to explain uncertainty and evidence transparently.
- •Different ‘conspiracies’ vary massively in plausibility (number of unlikely steps)
- •Quantitative thinking: more moving parts means lower overall probability
- •Experts should ‘show their work’ and communicate evolving uncertainty
- •Politicized institutional branding drives coalition-based rejection
- 41:02 – 42:15
Wrap-up: Pinker’s next project on common knowledge
In closing, Pinker previews his forthcoming work on the psychology of common knowledge—recursive layers of ‘I know that you know…’—and how it shapes social, political, and economic behavior. Chris ends the episode with thanks and channel sign-off.
- •Common knowledge: shared awareness and its recursive structure
- •Intuitive sense of what is public, irreversible, or ‘out there’
- •Applications across psychology, economics, and politics
- •Episode closing and sign-off
