We forecast the future using a broken simulator

Gilbert's foundational claim is that humans are the only animals capable of consciously simulating events that haven't happened yet, and we rely on this ability constantly to make decisions—choosing a spouse, a job, a vacation. But he argues the simulator has a critical flaw: it's fast, convincing, and almost always running, yet it borrows heavily from present-moment feelings and incomplete information rather than genuinely modeling the future.

He compares this to how a flight simulator can train a pilot only if it accurately represents real flight conditions; our mental simulator, by contrast, quietly substitutes shortcuts for accuracy and never flags the substitution. We feel like we're previewing the future when we imagine, say, retiring to a beach town, but we're really constructing a thin sketch filled in with assumptions we don't notice we're making.

The danger isn't that we imagine badly—it's that imagining badly feels exactly like imagining well, so we trust these forecasts with a confidence they haven't earned.

Takeaway: Treat your gut prediction about how a future event will feel as a hypothesis, not a fact.