At $50k per car (double the current average purchase price), the manual to autonomous transition would cost $900 billion. For America alone. That’s twice the cost of building the United States interstate highway system.pic.twitter.com/jnUZ3dk2Kh
You can add location information to your Tweets, such as your city or precise location, from the web and via third-party applications. You always have the option to delete your Tweet location history. Learn more
Cars will use cameras, lidar, and radar to see the next 200m, but need external information to know what is happening kilometers away. Road owners can provide this by using phone triangulation and video detection.pic.twitter.com/29SgbRBwO2
Cars will also need to know what other cars are doing and intending to do. This requires a communication channel (like Wifi-P) and information standard (yet to be defined).
My bet is on 1-way (think UDP), as opposed to 2-way (like TCP), data transfer between vehicles as well as with the road. No manufacturer wants a crash “because the road didn’t respond” or “the other car gave me the wrong information.” Cars need to be as independent as possible.pic.twitter.com/TKgnIbhzN5
On-demand cars on smart roads also means every person’s trip history will be stored, for as much as sharing platforms require identification.
These self-driving cars will also religiously follow traffic rules, as it will speed up the overall system and governments will make it part of their certification process.
No need to speed anyway, as self-driving cars react faster than humans and can thus go faster.pic.twitter.com/r28cPHdkz2
This does bring us to the point of regulation. Many observers point out that it might be the single largest stumbling block for full autonomy, though I’m hopeful.
People and companies in your place will see the personal benefits of self-driving cars once a neighboring city, state, or country allows for it. That creates competitive pressure to update regulation. Governments are slow, but this will move faster than you think.
Governments will want to certify that self-driving cars are “safe,” which honestly will be an industry itself. Most of these cars will use some form of machine learning, so we likely need simulations as opposed to formal verification.
Once regulation is adjusted, there will be a transition phase with mixed manual/self-driving traffic. This will make self-driving look worse than it really is, because the manual traffic is less predictable. We honestly can’t do much about it (see investments needed above).
Trucks will be the first ones to go autonomous, because their usage can be limited to the relatively easy confines of highways and companies are willing to put up with those limitations.pic.twitter.com/C0QdEUyOpf
Picture container depots at city edges and self-driving trucks shuttling between them. An extension of what’s already happening in ports.pic.twitter.com/RIXY8Mj6We
Over time, cars will be able to navigate the more complicated inner cities with their pedestrians, bikes, and other traffic. This is where things will get really interesting.
For one, more people will start using cars for trips than ever before, as the price will be low compared to other transport modes. We will run headlong into the Jevons paradox of improved efficiency causing usage increase. (Which makes the above $900b investment conservative)pic.twitter.com/cYY7MynviU
A big turning point will be once self-driving cars are cheaper than mass rail transport (tram, metro, train). The average per mile cost in the UK is £0.31 for cars vs £0.16 for trains now, so we need a 50% drop.
In the Netherlands, 10% of commuters take public transport and 50% go by car. In the United Kingdom, 15% go by train/bus and 70% by car. Roads will have to support 20% more cars if everyone would switch.pic.twitter.com/L5bi0iiRQG
The transport of choice for commuting changes dramatically if you look at cities instead of countries. In New York City, 60% go by public transport and 30% by car. In London, 50% take public transport and 30% go by car.pic.twitter.com/nq9BdZhnDo
Sharing platforms will likely get 30% of people to pool during peak hours through price incentives. Self-driving cars will also be able to drive much closer, more consistently, and faster. Which probably increases road throughput with another 50%.
The 30% carpooling and 50% road throughput increase should more than offset the 20% inter-city car usage increase. And cities will need a bit more work. Long live the roads we already have!pic.twitter.com/DSeBAEGwu2
The majority of public transport needs will be covered by car sharing platforms. Though there will still be a role for governments to incentivize coverage in low utilization/high-cost areas (i.e., rural).
Car sharing platforms can cover off-peak hours with regular capacity, and peak hours through carpooling and supply positioning. Demand super peaks, like natural disaster evacuations, are a challenge though. Will that be solved through capacity redistribution between areas?
Besides people switching from other modes of transport to cars, we will have new use cases for car transport. No more shuttling teenagers to football games. Elderly can see their grandchildren more often.
It’s also a question what people will do during these autonomous trips. Current rail transport (tram, metro, train) is likely a good comparison — especially for carpool trips. Expect lots of smartphone usage.pic.twitter.com/gi9tnbaV0R
As cars get smarter and accident rates go down, we will be able to get rid of airbags and seatbelts. This would completely open up in-car design and likely trend towards minimal and utilitarian furniture, as leasing companies optimize for cost.
Current public transport is likely the closest comparison for future self-driving car interiors. Low-cost, easy to maintain, and space efficient. All the things fleet operators love.pic.twitter.com/YQl4bLckTJ
Something else that will change: Self-driving cars don’t need parking spaces in urban areas. The cars will either be driving on the road or at the edge of the city for charging/cleanup.
It’s clear human driven cars will disappear from the roads over time. They will likely go the way of horses: Rich people toys with limited access to public roads and expensive to insure.
We are far in enough to do another recap: (1) 50% of the road will be digital. (2) Each car will make decisions on its own, as an integrated system will be too slow (to build). (3) Regulation will be slow and then rapid, as countries compete. […]
[…] (4) Highway trucks will be the first autonomous vehicles. (5) Cars will outcompete rail transport on existing road capacity. (6) Carpooling will increase and car interiors will look like public transport. (7) Pricing will drive parking spaces to outside the city.
I’m excited about what the future brings. Cheap, safe, and clean transport is a win for us all and self-driving electric cars seem the closest we will get to that. This will be the biggest change in transport since trains (1810), safety bikes (1885), and manual cars (1910).
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.