Shaping the flow of ideas at the frontier of cognitive systems.
—
While most companies respond to AI’s cost savings by cutting headcount, Whoop doubled its workforce to 600 people. The fitness tech company isn’t confused about AI’s capabilities — it’s betting on what happens when execution becomes radically cheaper. The divergence between Whoop’s strategy and the industry consensus isn’t about optimism or caution. It’s about whether you design for scarcity or abundance.
When the cost of producing something drops by 10×, the rational economic response is not to produce the same amount with fewer people. It’s to expand the total volume of what you produce. This pattern has a name: Jevons Paradox, after the economist who observed that when coal engines became more efficient in the 1860s, total coal consumption didn’t fall — it exploded. Cheaper energy didn’t reduce demand for energy. It made previously uneconomical uses of energy suddenly viable. Railways, factories, and heating systems that couldn’t justify the fuel cost at the old price cleared the bar at the new one. The efficiency gain unlocked latent demand.
AI is doing the same thing to execution cost. When building a tool, testing a strategic bet, or analyzing a dataset becomes 10× cheaper, the constraint isn’t how much you can afford to build — it’s how much ambition you can generate. The backlog of projects that never got funded because the headcount cost exceeded expected ROI doesn’t disappear. It becomes executable. Companies cutting headcount are capturing a one-time efficiency dividend: the same output for lower cost. Companies hiring against falling costs are positioning for the expansion dividend: vastly more output at marginally higher cost. The expansion dividend compounds.
—
The evidence for where this expansion is happening clarifies who the new builders are. Shopify’s internal data showed that the fastest-growing adopters of AI coding tools weren’t engineers — they were finance, sales, and support employees. Domain experts who previously had to wait for a developer to build the tool they needed are now building it themselves. At Cursor, an AI-assisted coding platform, 35% of merged code changes (proposed updates submitted for review) are now created by autonomous agents running in parallel on 10 to 20 simulated cloud computers at once. The bottleneck isn’t execution anymore. It’s clarity about what to build.
This inverts the traditional assumption about who benefits from AI coding tools. The global developer population is 36 to 47 million. The population of domain experts — financial analysts, marketers, support leads, operations managers — is hundreds of millions. When execution becomes cheap enough that non-developers can build, the addressable market for new tools doesn’t grow incrementally. It grows by an order of magnitude. The scarcity shifts from technical skill to domain knowledge. A financial analyst who knows exactly what report they need but can’t code it is now more valuable than a generalist developer who can code anything but doesn’t know what the business requires.
—
The strategic error most companies are making is treating AI as a cost-cutting tool rather than a capacity-expansion tool. Cutting headcount locks in the efficiency dividend but forecloses the expansion dividend. It assumes that the total amount of valuable work your organization could do remains fixed, and AI simply lets you do that work with fewer people. But when iteration becomes cheap, the physics of strategy change. A company that could previously test three strategic bets per quarter because each required a dedicated team can now test thirty. The constraint shifts from execution bandwidth to decision speed — how fast you can generate clarity about what’s worth building.
Whoop’s bet is that the companies designed for abundance will outcompete the companies designed for scarcity. When execution is expensive, you optimize for efficiency: do less, do it well, minimize waste. When execution is cheap, you optimize for exploration: test more, fail faster, expand the surface area of what you attempt. The expansion dividend accrues to the organizations that treat falling costs as an invitation to higher ambition, not as an opportunity to shrink.
The deeper shift is philosophical. For decades, organizational design has been shaped by the assumption that execution is the scarce resource. You hire carefully, prioritize ruthlessly, and gate ambitious projects behind ROI hurdles calibrated to high labor costs. AI doesn’t just make execution cheaper — it inverts the scarcity. Insight becomes the bottleneck. The question isn’t whether your organization can afford to build something. It’s whether your organization can figure out what’s worth building fast enough to matter.
—
Source: “AI cut execution cost by 10x. The companies cutting headcount are making the most expensive mistake of 2026” — Nate’s Newsletter, March 14, 2026.
Text summarized and optimized using Anthropic’s models and reviewed by a human.