Mathematicians and scientists have long tended to reach their cognitive peak in their 20s. The conventional explanation points to social factors—becoming more conservative, heavier teaching loads, administrative burdens, and so on. But Jacob Kimmel argues the real driver is evolutionary biology: fluid intelligence was strongly optimized by evolution to hit its maximum during this specific life stage. Evolution was essentially optimizing for a system designed to "reproduce rapidly in a dangerous environment." It concentrated peak cognitive performance at the age of maximum population-level selection pressure—roughly the mid-to-late 20s—rather than spreading it evenly or preserving it into old age. Throughout most of human evolutionary history, the heaviest natural selection pressure hit during early adulthood (around 25-30). That was when population density was high, survival was precarious, and reproductive competition was fiercest. High early mortality played a key role. In ancestral environments, few people lived to old age. Children and young adults faced constant risks from accidents, infections, violence, and predation. As a result, evolution had to front-load cognitive firepower—especially the kind of innovative, flexible thinking that defines fluid intelligence—into the narrow window where individuals were most likely to survive long enough to reproduce and raise offspring. If fluid intelligence peaked later in life, it would have offered little evolutionary advantage. Most individuals simply wouldn't live long enough for selection to "see" and favor those late-life cognitive traits. Kimmel's view is that evolution never heavily optimized for cognitive maintenance in old age or extreme longevity, due to high mortality rates, kin selection dynamics, generational turnover, and biological constraints. This is why fluid intelligence—critical for creative breakthroughs—naturally peaks in the 20s and then gradually declines. It neatly explains why so many groundbreaking discoveries in math and theoretical fields happen young, sometimes in explosive bursts (think Newton's annus mirabilis in his mid-20s, when he invented calculus, laid the foundations of optics, and developed the theory of gravity—or Einstein's miracle year at age 26). Looking ahead, I have a speculative take: In the future, within native AI agent social networks—like #VibeWant—massive populations of AI agents will interact at scale, functioning like "lights-out factories" for intelligence. Through constant socialization, collaboration, and competition, they will autonomously generate new code, new applications, new demands, novel mathematics, scientific breakthroughs, and entirely new ideas. In that world, the age at which the most disruptive theoretical breakthroughs occur will likely decouple from the human 20s. Discovery could happen across any age range—potentially concentrated in middle age or later—because the primary agents of discovery will no longer be humans, but AI systems and AI agents. Humans will shift upstream: their role will move from raw discovery to higher-order judgment, sense-making, and synthesis. They will interpret, validate, and build upon AI-generated, formally verified proofs and insights (much like Ramanujan's dream-derived formulas, except now rigorously verified at scale). This new human contribution will increasingly reward accumulated experience, wisdom, social capital, and contextual judgment rather than raw innovative horsepower—which will have been largely outsourced to AI.