While I agree with some parts with that post, I think the conclusion is wrong. I believe the root of the problem are the wrong terminal values. "Smart" technology is not put there to really solve a problem, but to drive sales. Providing real utility for users is secondary, as long as the new feature has big marketing value. Therefore I see the problem not with engineers and designers, but with salesmen giving orders and marketing-driven development.
As for lessons for driving, I see only two ways - either do actually train drivers like they should be, just like pilots are trained, or automate them out completely. Current level of training required for getting a driver's licence is a complete joke, and the attitude of most of the people who took or are taking the exam is just plain scary.
As for lessons for driving, I see only two ways - either do actually train drivers like they should be, just like pilots are trained, or automate them out completely. Current level of training required for getting a driver's licence is a complete joke, and the attitude of most of the people who took or are taking the exam is just plain scary.