This sounds like you're comparing driving a car to flying an airplane. From my understanding, much like before flying, you need a lot of foundational knowledge to work with ML and be productive on your own. You could learn to drive a car, like you can learn to build a Rails CRUD app over a weekend, you'll be bad at it still but you can get to point A->B with little investment.
The barrier to entry for AI stuff is quite a high. And IMO will be limited to certain kinds of people who like working on the hard science /math stuff, and have a strong enough early education to learn advanced linear algebra and statistics (for example) .
> not sure if one could learn to takeoff and land a Cessna badly in a week long bootcamp
Yes, you can. Taking off is extremely easy, you just need to accelerate with the flaps down and pull up at around 90km/h (if I recall correctly, it has been many, many years since I last flew an airplane).
Landing is much more complex, but you can do decent landings[0] in normal weather conditions after a few classes. The course I did (as I said, many, many years ago) had 35 flight hours and I did it in a summer (2-3 classes per week, usually 1 hour of theory and 45 minutes of flying) and you were perfectly able to do the exam (and pass) after it.
[0] Of course, in the eyes of many pilots, a decent landing is one you can walk away from. A good landing is one where you can use the plane again.
The barrier to entry for AI stuff is quite a high. And IMO will be limited to certain kinds of people who like working on the hard science /math stuff, and have a strong enough early education to learn advanced linear algebra and statistics (for example) .