Prototyping with AI

Maybe you found my website because you want to take advantage of the brave new world that seems to have emerged as the result of the advent of AI (or really LLMs, large language models). You wonder if you can finally try executing on that idea that you have had for a long time but hat was never feasible due to cost.

The good news for you: it is true that AI makes it cheap to build something real enough to argue with. I use that to answer the expensive question early: is this idea worth building at all?

A prototype here is deliberately thin. It exists to test one assumption, that people will use it, that the workflow holds up, or that the data is good enough, and it is built to be discarded if the answer is no. That is the point, and I will say so plainly rather than let a prototype quietly become the product.

If the answer is yes, there are two honest routes from here: write the requirements properly and build it once, or have me build the production version with the prototype as the reference. If a prototype has already made that jump without you deciding to, that is a different conversation.