We are on the early stages currently; ignore what Tesla/musk says; in 10 - 20 years full level 5 autonomy will be common place.
In the 80's the Cray 1 supercomputer was made, now I have so much more computer power in my pocket its frankly ridiculous. And it's runs on milliwatts rather than kilowatts.
Honestly not convinced, we've been promised self driving cars for almost a decade and we unearthed way more problems than ways of getting closer. And honestly, there are soo many situations where even I'm not sure what the correct course of action would be, so I'm not holding my breath for an AI sourcing from our collective actions.
I'm super curious how they will handle inclement weather where lanes cease to exist. And drive thrus. How will it handle a drive thru fully autonomously?
GPS accurate enough to tell the car where the lanes are.
Drive thru wouldn't be too bad, it's essentially stop and go traffic. Plus, when full autonomy happens, drive thrus will become less common. You can just pre-order your food while the car is driving you to the restaurant. The restaurant will tell the car how long the food will take, and what parking spot to use. Then your car will tell the restaurant your ETA, and notify them exactly when you arrive. Food is waiting, so you just grab the bag and go.
Yeah the issue with predictions around things like autonomous vehicles is it assumes tech advancements will flow in one direction/domain. There are still serious limitations of current AI systems in regards to autonomous vehicles and the next breakthrough may not be one relevant to that domain.
Its certainly possible, but we've basically been on the current autonomous vehicle hype for about 10 years already.
I work in the automotive industry. I believe we could be there in 10 to 20 years, but I'm not convinced we will be there.
Specifically, because vehicle autonomy has been a big buzz word in the industry for a decade or so, and it's starting to lose its zing. And when buzzwords lose their zing, the money dries up and the industry moves on.
Things like speed-adjusting cruise control and lane-keeping assistance, for example, are trivial to implement from a technological standpoint and don't cost much to add. But they don't show up in too many vehicles, because consumers stopped caring. I worked on a trailer backup assistance feature in a 2015 pickup that added zero production cost, but very few vehicles implement anything like that. Not because they're not valuable features, but because the industry loses interest and moves on.
The automotive feature that boggles me most is 4-wheel steering (where the rear wheels can move about 10 degrees or so). I've driven a vehicle with this feature, and it's an absolute game-changer. And it doesn't cost that much to implement either. Too bad the big OEMs don't care, because once you've driven one, you want it on every vehicle ever. Sigh.
I agree the logistics industry wants it. Though I think there may be problems to be solved for autonomous deliveries outside of the driving aspect (that is, the driver does a lot more than just drive). So I wonder if it'll pick up more if some of the other roadblocks for that are resolved...
General maneuverability is the main benefit. The back wheels more closely follow the front wheels, so for situations like a curved drive-through (curbs on both sides) or even just making turns on roads with narrow lanes, it's easier to not clip the curb (or other vehicles) on the inside of your turn. And at high speeds, you can implement "crab walk" to make lane changes easier. It makes a big difference with towing as well.