When I first started this lab, and started exploring questions with AI-built prototypes, I thought that I’d try to explore at least 1 question per week.
I did that for roughly the first 3 months and learned some things in the process, including, but not limited to:
- I have a lot of questions. This is less of a learning and more of a validation. But once I stopped being constrained by my ability to answer a question, I kept asking them!
- It’s easy to prototype things based on any question. Given the question, you can frame anything in a manner that can lead an LLM to try answering it in the form of a protoype.
- But not everything should be answered with a prototype. Some questions don’t need a prototype to answer them, or maybe can be better answered with other things (e.g., a deep research prompt). Others things might not be worth prototyping, for one reason or another.
- Prototypes really only go so far. Prototypes are a good start. And because they’re “cheap”, they can be discarded once they’ve served a purpose in scratching an itch. But once something is proven interesting, and you want to take it further, there’s a big gap between a prototypes and a real product.
So I no longer am trying to explore a brand-spanking new question every week. I do still plan to cultivate this habit of exploring new questions regularly, just not weekly. But in parallel, I want to go deeper on some of these projects, and turn them into full blown products.