An idea nobody acts on earns nothing and proves nothing. It survives as an anecdote, usually told with a nod towards the company that later did roughly the same thing and got paid for it. Value arrives with whoever takes the thing on.
There is a boast that comes round often: the speaker is bursting with good ideas. Dutch keeps a line ready for that posture. De beste stuurlui staan aan wal. The best helmsmen stand on the shore.
That boast got more expensive to make, and two days of actual work is now enough to settle it.
What an afternoon of prompting buys
Testing an idea used to be hard. Making one visible cost money and months. It costs an afternoon now. One argument doing the rounds holds that AI tools have made generating ideas, and even prototypes, trivial. It expects 2026 to be judged on execution and real results rather than on demos. When a convincing mock-up takes an afternoon of prompting, everybody ships the mock-up.
What stayed expensive sits behind the screenshot. The second version. Pricing that survives contact with a procurement department. The support inbox at eleven at night. The customer who renews without being chased.
Are good ideas scarce?
A mentor who has taught in MIT, Harvard and IE programmes argues that the adage "ideas are worthless, execution is everything" is misleading. He cites Paul Graham and James Altucher on good ideas being scarce and valuable in their own right. He points to Nicholas Bloom's research on the declining returns of new ideas. His conclusion: distance is only covered by executing a good idea well, and execution alone gets nobody far.
Fair enough. Effort poured into a bad idea also fails, often more expensively, because admitting it takes longer than starting it. Defending execution does not require pretending every idea is equally good.
A sharper version comes from an author who describes himself as an idea guy. He still holds that execution matters more for most startups, with one exception. Where the idea itself is the core of the value, the advice is to patent it and tell the world afterwards. That exception describes the mechanism rather than escaping it. A patent is a filing with a date on it, a fee, a specification written to survive an examiner, and years of willingness to defend it. Scarce insight turns into capital at the moment somebody converts it into an asset. The same author aims his warning at founders who lean on the idea and wait. The Field of Dreams school of business planning, where the thing gets built and they come.
The flattering half of that work is now available on demand, and priced accordingly.
What do two days settle?
The short brutal test is older than the tools. A 48-hour version circulates for exactly this purpose, on the reasoning that before spending serious time or money building a business, it is worth finding out whether there is something there to build (3BL Media).
So the principle is not new. The price of running it is.
Two days used to buy a few phone calls and a spreadsheet. Two days now buy the research and a working site, plus a first indication of whether anyone cares enough to click.
Speed is not value. AI will materialise a bad idea exactly as fast as a good one, so a weekend build proves only that building has become easy. That is correct. A site is not demand. But two days was never offered as proof that the idea works. It is proof that the person holding the idea is willing to find out. That willingness used to be hidden behind cost. It cannot hide there now.
Which leaves the people bursting with good ideas in an awkward spot. The excuse they had, the budget and the developer, is now worth about an afternoon. Spend the two days or decline them. Decline, and the idea stays exactly what it was at the bar: a story about a ship, told from the shore, by someone who has never been wet.