with reduced autonomy
I think the tech industry (specifically silicon valley) suffers from an industry-wide Dunning-Kruger effect - and I think its dawning on people now that we are close to 10 years down the line on autonomous vehicles.
"self-driving" is orders of magnitude more difficult than they understand....I mean wouldn't we first have things like self-driving trains or airplanes? Both of those are much simpler than the chaos of cars on roads....
good point, i do wonder why many trains still have conductors today. it seems like that is the easiest “self-driving” problem to solve.
reply
211 sats \ 2 replies \ @freetx 24 Jan
I think we are far more likely to see the "Self Checkout Cashier" model for trains / planes / cars. That is like Walmart has....1 human cashier overseeing 10 self-checkout kiosk....will be the same: One human driver remote-control driving 4 different trucks each with AI assist.
Or conversely maybe the model becomes: AI assist driving with lower-qualified driver behind wheel who can press a big red STOP button if things go awry.
reply
interesting, i think the supervisory role is probably most relevant for trains. wouldn’t that still be super difficult for one person to monitor 4 planes/cars at the same time?
the benefit with trains is that there aren’t many external environmental variables to consider.
reply
211 sats \ 0 replies \ @freetx 24 Jan
Well I think it could be arranged so out of the 4 trucks, only 1 could require active supervision at a time. Like 2 could be going down empty interstate, 1 is being unloaded/loaded somewhere, and 1 with active remote driver in city.
Also even if the truck needed to pull over on side of road for 20 mins until the human could take control, that would still probably be a big win compared to labor cost of 4 humans.
But yeah, trains seem like the slam dunk (but I don't know anything about the complexity of driving trains)
reply