You've got a lot of answers already but ... wow...
Allow me to say first of all that I agree we should have denser cities/towns, better public transport to get around, and we should all use it more.
But some stats (no citation, these are rattling around in my memory, feel free to fact check).
90% of cars are not utilised 90% of the time. That is insane wastage for energy intensively manufactured, expensive pieces of equipment.
1/3 of city centres are dedicated to parking (probably a US stat). Huge amount of valuable land unlocked.
Imagine unlocking that capital and land inefficiency. Then imagine cars halve in size and drive bumper to bumper and twice the speed. That could be a 2 x 2 x 2 improvement in commuting.
Don't get me wrong, even if we grant the pessimists on FSD that it won't be till 2100, that's a huge prize to be aiming at.
I don't understand the concerns over vehicle utilization and parking. Even if autonomous vehicles are made to work they will still be idle most of the time in order to have sufficient peak time capacity. Across most of the country, parking isn't really an issue because there's plenty of space. It's only a concern in a handful of dense cities.
There's no need to imagine, what you're describing already exists in Europe. We already know how to do this with today's tech, no need to wait for anything.
It's as simple as public transport + bicycles + dense mixed use zoning.
> Then imagine cars halve in size and drive bumper to bumper and twice the speed.
> Then imagine cars halve in size and drive bumper to bumper and twice the speed.
And now imagine one of those cars hitting a large pothole
Beyond this, how will there ever be 100% adoption to support that dream? There will always be a desire to have a steering wheel on some cars to go off road. It's really hard to see how that 2x2x2 dream could ever happen.
> And now imagine one of those cars hitting a large pothole
Or imagine the first time a self-driving car encounters a road-obstacle it flags it so that it can be repaired and all other cars avoid it automatically.
> There will always be a desire to have a steering wheel on some cars to go off road
Very true. People also own horses and burn candles, but these are not our primary means for transportation or lighting.
> it flags it so that it can be repaired and all other cars avoid it automatically
Doesn't help if you already have a multi-car pile up collision. OP said bumper to bumper at high speeds. Anything goes wrong and physics take over.
> Very true. People also own horses and burn candles, but these are not our primary means for transportation or lighting.
Horses very much still appear on roadways today. The idea that 100% of cars will be AI in the next 50 years I don't think is based in reality. Let's say we somehow get to a point where all cars have a switch to go from AI to not; so for example when on a highway where AI would shine, drivers could switch. What happens when a driver does not switch on the AI system? This is all to say the idea of that bumper-to-bumper driving at 100 mph with very small cars is snake-oil.
> but these are not our primary means for transportation or lighting
Not our primary means, but still a means and used. Again, this goes to the point, if you want that massive efficiency of AI cars, you need 100% adoption, not a single car in a convoy can be human operated. Alternatively, if the AI cars are smart enough to realize that another vehicle is human operated - do the AI cars slow down the entire convoy and then create adequate following distance? At that point, is that _any_ different from driving today?
> Or imagine the first time a self-driving car encounters a road-obstacle it flags
Perhaps the most common road obstacle is falling rocks. They shift around, there are many of them. They are 'repaired' by the bulldozer coming out and moving them all to the side, there is no guarantee that a human won't come out when it is quiet and move the obstacle. So, what is the use to even flag them?
How does an AI car know the difference between a small pothole and a larger one, and one that is even larger that needs to be avoided right now?
Back to the convoy example, let's say a tire blows; the whole convoy then crashes. It's just too easy to picture so many scenarios where a convoy of bumper-to-bumper cars turns into one part junkyard, one part graveyard.
We will get to a point where it becomes unethical to let humans drive cars (outside hobbyist places). A Sam Harris says, "the ape behind the wheel" just causes issues.
"Just causes issues" is discounting what humans are good at and discounts the idea that AI technology in cars could complement humans. Rather, that point of view I think is most useful if you are trying to get to 100% AI cars so that you can sell more roadways and more cars. Peak-hype sir, peak hype.
I'm reminded of the promises from those 1950s propaganda videos for building more highways. That reality turned into the gridlock of LA freeways. [1]
Who cares if a car is only utilized 10% of the time. At 10% use the car is going to last an average of 20 years. Use it 90% of the time and it’s gonna get junked in 1-2 years.
Allow me to say first of all that I agree we should have denser cities/towns, better public transport to get around, and we should all use it more. But some stats (no citation, these are rattling around in my memory, feel free to fact check). 90% of cars are not utilised 90% of the time. That is insane wastage for energy intensively manufactured, expensive pieces of equipment. 1/3 of city centres are dedicated to parking (probably a US stat). Huge amount of valuable land unlocked.
Imagine unlocking that capital and land inefficiency. Then imagine cars halve in size and drive bumper to bumper and twice the speed. That could be a 2 x 2 x 2 improvement in commuting.
Don't get me wrong, even if we grant the pessimists on FSD that it won't be till 2100, that's a huge prize to be aiming at.