This is not a really strong argument against self-driving cars. The fact that a problem is NP-hard doesn't make it untractable. Every day we use apps that deal with NP problems (e.g., routing problems, packing problems, etc.). Also, note that there're P problems whose instances can be harder than (smaller) NP ones.
My steps for learning French and Spanish were:
1) 1/2 months nonstop on Duolingo
2) Cartoons/books for kids
3) Spending free time with French/Spanish speakers
(gotta mention that I was lucky because my native language is romance and I lived in both France and Equador)
"The Art of Computer Programming" by D. Knuth taught me the beauty of well designed algorithms. "The Art of Prolog" by E. Shapiro opened my mind to new programming paradigms.
Jaron Lanier's "You Are Not a Gadget" has some nice thought about this point, claiming that the web in the 90s was a way more creative and personal space than today.
SOCML (Self-Organizing Conference on Machine Learning) is really nice. Open discussions and interesting people. It's not the classic academic conference.
Murano's glass blowers (Venetian lagoon) are the craftsmen that I admired the most. It's a pleasure seeing them at work. Immense skills and tradition. https://en.wikipedia.org/wiki/Murano_glass