There's space for what I call bottom feeder ideas. These are the multimillion dollars that VC money won't fund because their model needs a billion dollar hit to break even. Most of these can't hire teams of people.
This could be a better hardware store, some kind of directory, better notekeeping, some kind of text based dating app. I made good money off a recipe app. A lot of the world still runs on spreadsheets, when they shouldn't have to. The key to this is to have a niche enough interest, to get to a point that you see something that should obviously be solved with tech but isn't.
AI opens up a lot of opportunities for automating routine mental tasks. Robots did a good job of automating manual tasks. We still have things like ghostwriting and brainstorming names and designs. Robots do not assemble 100% of a car or chocolate bar, but they do a lot. Right now, AI is doing 0% of some of these, when there's plenty of potential to do 20% or even 70% of the work. People get too hung up on getting it to do all the work. AI won't replace comedians, but it will help them write jokes.
You also can't write a good joke-writing AI unless you have an existing process for writing jokes. So you need to somehow be in the top 20% for comedy writing and 20% of AI development to do this. That's quite a niche. And also not something you can outsource to a team.
