The entrepreneur literature documents a genuine discontinuity in startup economics that most investment frameworks have not yet absorbed. The core claim — that a solo founder with domain expertise and access to the right AI stack can compress what once required a twelve-person team into a one-person operation capable of reaching revenue before a second hire — is empirically grounded. YC’s Winter 2025 batch showed 25% of startups with codebases that were predominantly AI-generated. Lovable crossed $100M ARR in eight months. Cursor reached $500M ARR with revenues doubling every two months. The productivity ceiling for the competent AI-augmented founder has moved dramatically.
But the two core findings contain a tension that neither fully resolves. The solo founder thesis argues that deep domain expertise is now the decisive founding advantage — that non-technical founders who understand a problem space intimately can build production software without an engineering co-founder. The AI-native software development evidence documents that AI-generated code contains 1.7 times more major issues than human-written code, 2.74 times higher security vulnerabilities, and an industry-estimated 8,000 of 10,000 production builds requiring expensive rebuilds. The democratization is real. So is the fragility.
The resolution is probably this: AI coding tools have genuinely lowered the bar for reaching a functional prototype, but they have not lowered the bar for building something defensible and scalable. The “vibe coding to production” arc works until it does not, and the reckoning typically arrives at 90 days when scaling exposes the technical debt accumulated during fast iteration.
For investors, the implication is that lean founding teams are no longer a signal of execution risk — they may signal superior leverage. But the due diligence question shifts: not how many engineers, but how much does this founder actually understand the system they are building. That understanding is not evenly distributed, and its absence does not show up until it matters.