Applying Jevons Paradox, as AI makes code (the "digital brick") cheaper and easier to produce, it doesn't reduce the need for builders. Instead, it unlocks previously uneconomical or overly complex projects, thereby increasing the overall demand for software engineers to build more ambitious things.
Silicon Valley insiders building AI may overestimate its impact due to self-interest (looming IPOs) and a narrow perspective. Their expertise in AI doesn't translate to economics or labor markets, and their track record of understanding the world outside their bubble is poor, making their job apocalypse predictions unreliable.
Stripe Atlas, a service for new company incorporation, saw a 130% year-over-year increase in Q1. This data point, combined with observations that AI startups are growing revenue faster than historical norms, provides tangible evidence that AI is lowering barriers to entry and sparking a significant wave of entrepreneurship.
Fears of an AI investment bubble are contradicted by market data showing that customer backlogs for cloud capacity are growing significantly faster than the massive capital expenditures by providers. For example, Mag7's Q1 backlog was $1.3T against $400B in spending, indicating that current investment is driven by real, committed demand, not just speculation.
Initial AI market skepticism was based on a SaaS model of selling limited-value subscriptions ('seats'). The new reality is a utility model based on consumption ('tokens'). In an agentic era, a single user can drive thousands of dollars in token usage, creating a virtually uncapped revenue stream that justifies massive infrastructure investment.