Thinking, Fast and Slow — Kahneman

Daniel Kahneman’s 2011 synthesis of decades of behavioral economics research. Dense but worth it. This is an opinionated summary — I’ve kept what I use and cut what I don’t.

The Core Model: System 1 and System 2

System 1 — Fast, automatic, unconscious. Pattern-matching, emotional, associative. Operates continuously. Generates the “feeling” of intuition.

System 2 — Slow, effortful, deliberate. Handles logic, calculation, self-control. Lazy by default — it delegates to System 1 wherever possible.

Most errors in judgment happen when System 2 should override System 1 but doesn’t — either because we’re tired, rushed, or the System 1 answer feels correct.

Key Biases Worth Knowing

Availability heuristic: We judge probability by how easily examples come to mind. Plane crashes feel more common than car accidents because they’re more vivid in memory. Corrective: ask “what’s the base rate?”

Anchoring: The first number we hear disproportionately influences our estimates. In salary negotiations, in price appraisals, in charitable giving. Corrective: generate your own estimate before hearing the anchor.

Halo effect: One positive trait bleeds into our assessment of everything else. Attractive people are judged smarter; early wins make later failures less visible. Corrective: disaggregate assessments; evaluate attributes independently.

Planning fallacy: We systematically underestimate how long tasks will take and how much they’ll cost, while overestimating benefits. Corrective: reference class forecasting — look at how long similar projects actually took.

Loss aversion: Losses hurt roughly twice as much as equivalent gains feel good. This is why people hold losing stocks too long and sell winning ones too soon. Corrective: reframe decisions in terms of final states, not gains/losses from current position.

What Kahneman Is Confident About

• Human judgment is systematically biased in predictable ways
• Expertise is real but domain-specific; experts are not immune to cognitive biases outside their expertise
• Statistical thinking does not come naturally; it requires deliberate effort
• Simple algorithms often outperform expert judgment for prediction tasks

What the Book Doesn’t Claim

Kahneman doesn’t argue System 1 is bad or that you should always slow down. Intuition trained by genuine feedback loops (chess, firefighting, medicine) is often reliable. The problem is knowing when your intuition is trained vs. when it’s confabulation.

Practical Takeaways

1. Slow down on irreversible decisions. System 2 is expensive; deploy it where the stakes are high.
2. Pre-mortem any important plan. Before committing, ask: “It’s a year from now and this failed. What went wrong?” Forces System 2 engagement and counters optimism bias.
3. Separate idea generation from evaluation. Anchoring means the first idea on the table shapes all subsequent ones. Generate options independently before discussing them.
4. Use checklists for recurring high-stakes decisions. Externalise System 2 reasoning; don’t rely on willpower.
5. Track your predictions. You can’t calibrate intuition without feedback. Keep a decision journal.

Book: Thinking, Fast and Slow — Daniel Kahneman (2011, Farrar, Straus and Giroux)