The Power Law of AI: Why Averages Don't Matter Anymore

AI is accelerating a shift from “averages” to “power laws,” where small differences in skill, execution, or product fit can produce outsized rewards. In a traditional world, performance and outcomes cluster around a norm—so hiring, software procurement, and career progression often revolve around being “better than average.” In the new reality, rewards and opportunities concentrate at the top: the top 1% in a job family can capture disproportionate gains, while the rest increasingly feel stuck below average even if they’re objectively competent.

That power-law pattern shows up across careers, talent markets, and product distribution. In workplaces, compensation and advancement don’t follow a smooth, normal distribution. Instead, extraordinary ability—whether in product management, engineering, or technical leadership—tends to compound into dramatically larger outcomes. The same dynamic applies to builders: success isn’t about who “started” a project or how impressive the concept sounds; it’s about whether the product works, solves a real problem, and communicates value clearly enough to win in the marketplace. When AI enters the picture, it acts like an accelerator that amplifies tiny advantages. If someone can prompt, iterate, or execute slightly better than peers, they can produce significantly more work—making the gap widen over time as models improve.

Product distribution follows the same logic. A founder who lands in the top tier of distribution can generate breakout results quickly, not necessarily through cheating or unfairness, but because the system reinforces top performers. AI exaggerates small disparities in capability and turns them into large differences in output and traction. Over time, these advantages stack as “power curves” steepen—so a product that leverages AI correctly, evolves with market needs, and fits the right use case can achieve rapid, even overnight, breakout success.

The practical takeaway is career strategy: treat the environment as a power-law world and plan accordingly. Moving up the curve yields nonlinear benefits. Improving from the 40th percentile to the 70th percentile can deliver far more than a simple 30-point gain, and pushing from the 90th to the 95th percentile can be even more valuable. Luck and networking matter, but the transcript emphasizes “time on station”—persistent, obsession-level focus on a subject or craft. That sustained effort builds connections that become networks and unlock options.

For individuals and teams, the advice is to pick a specific “power curve” to ride—whether in product management, engineering, or a narrower niche—and invest in the trajectory that leads toward top-tier outcomes. For companies, the same lens applies to product strategy and adoption: failures often come from not grasping the exponential nature of the new problem space. The message is blunt but consistent: power laws and AI are here to stay, and the only controllable lever is where to position oneself on the curve—because assuming the world still runs on averages can leave people far below where they could be.