Over a long period of time I worked with computers and technology doing research and development type things, and over the duration I met many "ideas people". I never liked them, and they always turned out to be talentless tossers. Having ideas is easy. It happens to everyone all of the time. Except when asleep, perhaps. But implementation of the ideas, and the specifics of that, matter far more. Sometimes realising an idea in the real world can take years, or be the work of a lifetime.

The ideas people were always "you should just". "You should just get a job in...". "You should just have the system do such-and-such". "You should just have world peace". They were hand waivers, never concerned with the specifics because that was for the people beneath them in the chain of command to figure out the details. Notice also the similarity with waterfall development methodology. Linear. Heirarchical. Presumptive.

Since in the past I did stuff with robotics I would sometimes encounter the "ideas people" who would insist that the machine "should just do" something which I knew was outside of the known laws of physics. They thought they were being dynamic, entrepreneurial innovators pushing the boundaries.

But no, they were just tossers.

The thing is that the world is always far more complex than your mental model of it. It takes a lot of interaction and itteration to both understand aspects of the world and to be able to try to change it. Especially in technology, the specifics matter a lot and can't just be hand waived over. If you are doing anything even vaguely important, you will often be humbled by how wrong your previous mental model of the world was, when actually put to the test.