Here's my negative scenario: self-driving is "AI-complete"; you can't really hit all the edge cases without solving AI in general, which is more than 30 years away (Kurzweil is the wild optimist and predicts 2045).
You CAN use self-driving in limited circumstances, but those limitations are precisely the ones that make driving yourself around more attractive. The expense doesn't go down as quickly as anticipated because of this. They are a niche technology for DECADES.
Just like 3D printing -- the tech obviously works to an extent, but it's just not appealing for many use cases. Same with VR -- the tech works to an extent, but the content is not appealing to general audiences.
I'm happy to be proven wrong though, because I would be the first one to buy a self-driving car where I could fall asleep behind the wheel. I would love to take a trip from SF to Portland or LA like this.
> because I would be the first one to buy a self-driving car where I could fall asleep behind the wheel. I would love to take a trip from SF to Portland or LA like this.
You likely wouldn't be buying a car anymore when this becomes possible... this entire video series is about the replacement of individual car ownership with fleet based systems.
This is the big thing people don't factor in:
- the AI "problem" gets easier and easier the more cars talk to each other
- the average person won't need to 'afford' self-driving cars or buy add-ons for their cars, they'll just use Uber/Waymo and save money
- mass produced smaller fleet cars designed specifically for the job reduce costs per car
- fleet subscription prices to become cheaper and cheaper than owning a car (via competition, technical progress, etc), further increasing adoption
- 10x more developers/designers/traffic engineers start focusing on the problem (with real life/death consequences)
- 100x more real-time data is shared, thousands of new unexpected events become optimized each day
- AI cars get safer and safer, creating great social and altruistic pressure to not drive non-AI cars
- creating public health incentives as less traffic fatalities and law enforcement reduce expenditures
- countless businesses get wealthy off it (Waymo/Amazon/etc) and reinvest in the industry and R&D
- hundreds of billions in capital gets poured in rather than tens of billions
- traffic/infrastructure begins to be designed around them, dedicated lanes are built, further increasing the utility of using fleets/self-driving > manual
Once it hits a certain inflection point of adoption and utility it will be explosively exponential rate of adoption, just like smartphones.
Even if self-driving is AI complete , a taxi that knows how to self drive in the highway and easy areas of the city, but is remotely operated otherwise is extremely useful.
Add the ability to remotely drive a platoon of cars and it's even more interesting.
I think it will halve as the tech progresses, just due to optimization. Also, no windshield means better aerodynamic possibilities that could compensate for the loss.
Self-driving trucks that make 95% of the route on highways automatically, and only take a driver for a couple last miles and the loading-unloading operations would still be a huge disruptor.
A high-profile Uber investor has said that level 4 self-driving won't be in the U.S. for 25 years:
https://www.cnbc.com/2017/04/06/bill-gurley-uber-investor-se...
That means they can service people without driver's licenses and people who can't drive -- the old and the young.
Also, Chris Urmson said the transition will take 30 years:
https://spectrum.ieee.org/cars-that-think/transportation/sel...
https://mondaynote.com/autonomous-cars-the-level-5-fallacy-2...
Here's my negative scenario: self-driving is "AI-complete"; you can't really hit all the edge cases without solving AI in general, which is more than 30 years away (Kurzweil is the wild optimist and predicts 2045).
You CAN use self-driving in limited circumstances, but those limitations are precisely the ones that make driving yourself around more attractive. The expense doesn't go down as quickly as anticipated because of this. They are a niche technology for DECADES.
Just like 3D printing -- the tech obviously works to an extent, but it's just not appealing for many use cases. Same with VR -- the tech works to an extent, but the content is not appealing to general audiences.
I'm happy to be proven wrong though, because I would be the first one to buy a self-driving car where I could fall asleep behind the wheel. I would love to take a trip from SF to Portland or LA like this.
Another good link: https://medium.com/self-driven/a-decade-after-darpa-our-view...