Iregbulem, a Lightspeed investor focused on developer tools, argues that unlike physical goods where copies retain value, software is an intangible good where only net-new output matters, meaning developer productivity should be measured by novel contribution rather than gross code produced, and that idea-generation productivity mathematically tends to decline over time in any innovation-driven field. He supports this with Bloom, Jones, Van Reenan, and Webb's semiconductor research, showing chip researchers grew roughly 18-fold between 1971 and 2014 just to sustain Moore's Law's steady rate of transistor density gains, implying research productivity fell to about 5% of its earlier level, and argues software engineering follows the same idea-scarcity dynamic as any research-driven field, meaning developer productivity naturally erodes as more has already been built. To counteract this decline, he proposes a four-stage developer productivity flywheel: better tools increase individual developer output despite an initial adoption dip, that higher productivity pushes companies to hire and pay more engineers, more productive engineers collectively ship more software including new productivity tooling itself, and that tooling feeds back into stage one, arguing no single fix exists and the flywheel must be driven from every possible angle at once.

Investors building a developer-tools thesis get a rigorously argued, economically grounded case for why this category deserves dedicated attention, since the semiconductor R&D parallel offers a genuinely testable, historically documented mechanism for declining productivity rather than relying on anecdote or a vague sense that software has become harder to write. Founders building developer productivity tools themselves get a clear articulation of the specific market dynamic their product addresses, and a concrete distribution insight, that thoughtfully applied tools produce an initial adoption dip before net gains appear, useful for setting customer expectations and choosing rollout pacing rather than promising immediate productivity gains that a J-curve effect would undermine. The four-stage flywheel model, tools to hiring to output to more tools, gives operators and investors a shared mental model for how gains in this specific category compound over time, useful for evaluating whether a developer tool is likely to produce a one-time efficiency bump or a genuinely self-reinforcing growth loop within an engineering organization.