The Gap Between What We Know and What We Decide

In September 2025, NASA's Perseverance rover found what might be the most significant thing humanity has ever stumbled across: small reddish spots on Martian rocks, ringed by dark material, carrying organic molecules, formed in conditions that on Earth almost always mean biology. The scientists called them leopard spots. The cautious official language was that they represented the "clearest sign" yet of ancient life on Mars. And then, more or less, the news cycle moved on.

That's not a criticism. It's a clue.

The leopard spots story illustrates something quietly strange about this moment in history. We are accumulating discoveries, warnings, and data points at a pace that our decision-making machinery was never built to process. The knowledge arrives faster than the response. The gap between what we know and what we actually do about it keeps widening, and that gap has a shape that behavioral economists would recognize immediately.

Daniel Kahneman spent decades mapping the architecture of that gap. The core finding, underneath all the specific biases, is that humans are not bad at thinking. We are very good at a particular kind of thinking, the fast, pattern-matching, low-effort kind, and that kind of thinking is poorly suited to low-probability, high-consequence events that unfold slowly or abstractly. A leopard spot on a Martian rock 140 million miles away, hinting at life that died 3.5 billion years ago, does not trigger the same cognitive alarm as a car pulling out in front of you. The stakes are incomparably higher, but the urgency register stays quiet.

The ODNI's Global Trends 2040 report makes a related observation from a geopolitical angle. Technology timelines are shrinking, it notes, from decades to years and sometimes faster. Countries and institutions that once had the luxury of watching a technology mature before committing to it now have to choose before the implications are understood. The behavioral economics framing for this is bounded rationality: you don't make the optimal decision, you make the best decision available to you given your cognitive limits and time constraints. When those time constraints tighten, the quality of decisions tends to fall, even among smart, well-resourced actors.

So here is the uncomfortable intersection. The rate at which genuinely consequential things are happening, potential signs of past life on Mars, accelerating AI capability, materials science breakthroughs, is increasing. The time available to process and respond to those things is shrinking. And the cognitive hardware we bring to that challenge is the same hardware that evolved to track predators on a savanna.

This is where nudge theory, the practical toolkit of behavioral economics, turns out to be both more important and more limited than its proponents usually admit. Nudges work well for predictable, repeated decisions with clear feedback loops. Enroll people in pension plans by default and they save more. Make the healthy food more prominent at the cafeteria and people eat better. The feedback is fast, the behavior is habitual, and the intervention is low-friction. But the decisions that matter most right now, how to govern AI, how to respond to early evidence of extraterrestrial biology, how to allocate resources across competing civilizational risks, are none of those things. They are novel, they have no feedback loops yet, and their consequences play out over decades.

For those decisions, the nudge is the wrong tool. What you need is something harder to build: institutions and processes that compensate for the specific cognitive biases most likely to cause problems. Present bias, which is the tendency to weight near-term costs far above long-term benefits, is probably the biggest one. Loss aversion runs close behind. A finding that hints at past Martian life should, if we were processing it rationally, trigger a serious reorganization of scientific priorities. Instead, NASA faces mission cancellations and layoffs while the samples sit waiting.

None of this is hopeless. The fact that we can name these biases and trace their effects is genuinely useful, not just academically but practically. Some institutions have gotten better at building in the equivalent of pre-mortem analysis, asking not "will this work?" but "if this fails, what will have caused it?" That reframe tends to surface the slow, abstract risks that optimism bias buries. It is not a solution. It is a partial correction.

The leopard spots on Mars might turn out to be nothing, an abiotic process we haven't fully characterized. Or they might be the first hard evidence that life is not an accident peculiar to Earth. Either answer changes everything about how we see our situation. The odd thing is that we might already have the answer sitting in those rock samples, waiting for the right instrument to read it, while here on Earth the bigger challenge remains stubbornly human: building the decision-making capacity to handle what we might find.