Disposable Software: The Trend 90% of People are Getting Wrong

Disposable Software: The Trend 90% of People are Getting Wrong

More

Descriptions:

Nate B Jones argues that “disposable software” — one of the most discussed concepts in AI in early 2026 — is widely misunderstood, and that the misunderstanding is leading builders and companies toward bad strategic decisions. He defines disposable software not as a philosophy but as an economic inevitability: when the marginal cost of producing something collapses toward zero, that thing becomes disposable by default, in the same way digital photos became disposable once storage cost nothing. The evidence he cites is striking — 95% of the last Y Combinator batch shipped products with codebases that were over 95% AI-generated, Lovable hit $100M ARR in eight months, and Cursor’s valuation jumped from $2.6B to $29B in a single year.

The video’s central distinction is between two phenomena that share the label but are fundamentally different. The first is genuinely new: personal throwaway software — one-off dashboards, family trip apps, weekend games — a category that barely existed five years ago and is now trivially buildable with tools like Claude Artifacts or Lovable. The second is more contested: the claim that companies should vibe-code all their internal tools and potentially their core products. Jones dismantles this using attention economics. Cursor’s CEO built a working browser with AI agents running for a week, generating over 3 million lines of Rust code — but someone still had to decide to build it, configure the agents, evaluate the output, and maintain it afterward. Software cost was never the real constraint on what gets built; attention always was, and cheap software doesn’t make attention more abundant.

He also cites research showing AI-generated code introduces security vulnerabilities in nearly half of all coding tasks, typically at an architectural level that automated scanners and human reviewers struggle to catch. The video closes with a clear strategic segmentation: maximum disposability works for companies like Cursor, whose customers are developers and who benefit from moving fast and letting users adapt. For enterprise software companies with stable, regulated customer bases, the maintenance burden and attention cost of vibecoded internal tooling quietly offsets whatever was saved on initial development.


📺 Source: Nate B Jones · Published January 20, 2026
🏷️ Format: Opinion Editorial