I advise taking a look at the mir titles or the Israel Gelfand books. Looking from it, Russian are really good at producing top notch popular science.
Tough there is no fry, their math book are almost game like, with few carefully built example and very clear explanation using only some diagram when needed (consequently their book are quite small) and the exercise are absolutely not rote based, except the first few exercice, and even then they all serve to illustrate a specific part of a concept, all the others are puzzle like problem.
Putting lots of full color image is not making math "fun", well built and interesting problem is.
Maybe partner with the creator of the XXX simulator series, you could make some kind of "MMO" where each participants have to run errands the safest way possible while interacting with each other, and upload the training data. Some player could be randomly elected as "maverick" whose goal is to crash and cause accident, the other player would have to handle them.
And if driving in highway is a problem, why not use a test terrain complete with fog generator? With RC car representing pedestrian, other car, animal, ... feed the video first into an AR system and then give it to the neural net.
I totally agree, and this is why I am quite angry with the "basic is bad for your brain" statement from Djiktra. I understand the sentiment, but the idea that starting with very basic "non-rigorous" language is not only a waste of time but will even hold you back make me lost several years on my programming career.
As a budding a student of electronics, I found the mims book to be while not very rigorous to provide a good general overview and intuition on the field, which makes subsequent more "serious" more manageable, could you give more details as to why I should not read it?
Game that use hardware hack are also fascinating, the superfx chip story is very nice.