Proof Over Prototype

The threshold keeps rising

Geoffrey Woo's pitch principle for an era of generative AI and commoditized manufacturing: "the value of prototypes, wireframes, and ideas trends to zero."¹ He can "prototype any software idea in one weekend, and launch it the following weekend," and can get a Chinese co-manufacturer to clone a consumer product in weeks, so an app wireframe or a TestFlight alpha with a hundred friends and no retention data moves him not at all.¹ Product, in his framing, is necessary but insufficient, and "the proof threshold of your competency increments higher every day."¹

What actually counts as proof

Money and sales, plus customer acquisition cost, retention, and growth rates. Or, in Woo's words, "something technical that I cannot build in a weekend."¹ Plus repeatable systems for improving the product and a scalable path to distribution. Woo's reasoning is the fundraising-side version of the claim that there is no moat in software: "any good idea will have dozens of competitors crop up as soon as any traction is demonstrated," so defensibility has to come from something a clone cannot simply copy.¹ Because product itself is increasingly commoditized, distribution becomes the primary skill, and proof of it often starts close to home: showing that a founder's own friends and professional network are already paying customers, on the logic that "if you can't get your people to pay you, then why would you expect random strangers to buy from you."¹

The one exception in his framework is a genuinely new category of science or engineering, artificial general intelligence, fusion, interstellar travel, teleportation, immortality pills, where inventing the category at all maxes out the proof bar and the founder can skip go-to-market entirely: in Woo's telling, that founder is "on your path to be a trillionaire and you are a top 10 human of our time."¹

Why it matters

It resets what having something means inside a pitch. In a world where anyone can generate a working prototype in a weekend, the prototype itself is worth close to nothing, and the scarce thing becomes evidence that value is real and defensible. It is the operational consequence of the idea that a VC's job is to make money for its limited partners: an LP is underwritten on proof, not on vibes, so a founder's task is to lead with revenue, retention, and a hard-to-copy edge, a real wedge, a proprietary integration, or regulatory or relationship depth, rather than with a polished demo.¹

Tensions

At the earliest, pre-traction stage the rule risks a chicken-and-egg problem, since some categories, deep tech, regulated industries, long-sales-cycle enterprise software, cannot show customer acquisition cost or retention on day one, which is presumably why Woo carves out the science exception. Sometimes proof has to mean proof of the founder rather than proof of the product. And "I can build it in a weekend" is a claim specific to software and consumer hardware; it understates domains where the real moat is operational, capital-intensive, or trust-based, and cannot be built at all on that timeline.