3 ms·

I think this article actually makes quite a fresh and interesting point, after the first few paragraphs that sound like stock self-driving skepticism. The autho

I think this article actually makes quite a fresh and interesting point, after the first few paragraphs that sound like stock self-driving skepticism. The author is pointing out that imagining self-driving cars of the same form-factor and niche as the human-driven cars today is a fallacy on par with a 1990s futurist imagining the internet as simply a way to send letters faster. The point is that changes to the technological landscape as momentous as self-driving rarely are as simple as 1:1 replacement of the technology that existed before: the self-driving car is not unattainable, but it is a red herring. Given this, I don't really know why he mentions the challenges involved in self-driving and his skepticism thereof -- it seems to me that his autonomous "rovers" and "software trains" face basically all of the same challenges as a self-driving sedan would. Overall I applaud the imagination that the author applies, bringing together the threads of autonomous vehicles, micromobility, and lingering anxieties from COVID into a set of bold claims about the future of urban transit. What I am curious about is whether the future he paints ("micro-sprawl"; "walking is passé") is one people find compelling and desirable. I think he's basically right that the underlying market forces will lead us there in the next couple of decades absent strong action from municipal governments, but is this the right way to model cities of the future, or are we sleepwalking into another Robert Moses style disastrous reimagining of the urban fabric? Personally I'm undecided.

ppod 6y ago
Yes, I think a lot of people here didn't read beyond the first third of the article, which is understandable because it tries to be very poetic and sci-fi and fails to actually make clear what it's talking about until more than halfway through.
ImprobableTruth 6y ago
>"software trains" face basically all of the same challenges as a self-driving sedan would. How so? It seems pretty evident to me that having separate, dedicated lanes should make it a lot easier. What I'd say is by far the biggest concern, human-driven traffic, is just completely eliminated with that.
- AstralStorm 6y ago
  It's not, but because rail is much stricter and the safety protocols semi-automated to fully automated, the self driving is much easier, and unexpected conditions much easier to detect because expected conditions are very uniform. Self driving supervised trains can drive on non-ETS (not fully automated tracking and safety) rails too with additional precautions. Lines where trains are driven by humans, and where some safeties are manually operated, as long as the presence of those safeties and cars is programmed. These behave then more like driver assist and automated safety.
specialist 6y ago
Agree with your positive response. Missing from this conjecture is a synthesis with car and bike sharing. There will also be massive land use, zoning changes when there's 90% fewer personal cars. Unused capacity (parking, roads) freed up and repurposed. The role for autonomous driving for moving people (vs turking cargo) will look very different when our cities look more like medieval villas and college campuses. I still cannot even guess if the future will be more concentrated or more sprawl. So I've provisionally decided "both", but very different from today.
- ip26 6y ago
  90% fewer? To me this would seem to ignore that everyone tends to want to go to the same place at the same time. Demand for automobiles (and parking, and roads) is peaky.
  - specialist 6y ago
    90% is my optimistic guess. I'd wager today's style (cars, SUVs) will downsize even more. It'd be interesting to learn the automobile industry's own market size projections.
- Fricken 6y ago
  Most American cities are desperate for massive land use and zoning changes now, but I don't think it could happen without the implementation of broad mandates at the national level, such as you see in Japan. Politically that just seems impossible these days. Maybe it will seem more possible in a few years.

ppod 6y ago

Yes, I think a lot of people here didn't read beyond the first third of the article, which is understandable because it tries to be very poetic and sci-fi and fails to actually make clear what it's talking about until more than halfway through.

ImprobableTruth 6y ago

>"software trains" face basically all of the same challenges as a self-driving sedan would. How so? It seems pretty evident to me that having separate, dedicated lanes should make it a lot easier. What I'd say is by far the biggest concern, human-driven traffic, is just completely eliminated with that.

AstralStorm 6y ago
It's not, but because rail is much stricter and the safety protocols semi-automated to fully automated, the self driving is much easier, and unexpected conditions much easier to detect because expected conditions are very uniform. Self driving supervised trains can drive on non-ETS (not fully automated tracking and safety) rails too with additional precautions. Lines where trains are driven by humans, and where some safeties are manually operated, as long as the presence of those safeties and cars is programmed. These behave then more like driver assist and automated safety.

specialist 6y ago

Agree with your positive response. Missing from this conjecture is a synthesis with car and bike sharing. There will also be massive land use, zoning changes when there's 90% fewer personal cars. Unused capacity (parking, roads) freed up and repurposed. The role for autonomous driving for moving people (vs turking cargo) will look very different when our cities look more like medieval villas and college campuses. I still cannot even guess if the future will be more concentrated or more sprawl. So I've provisionally decided "both", but very different from today.

ip26 6y ago
90% fewer? To me this would seem to ignore that everyone tends to want to go to the same place at the same time. Demand for automobiles (and parking, and roads) is peaky.
- specialist 6y ago
  90% is my optimistic guess. I'd wager today's style (cars, SUVs) will downsize even more. It'd be interesting to learn the automobile industry's own market size projections.
Fricken 6y ago
Most American cities are desperate for massive land use and zoning changes now, but I don't think it could happen without the implementation of broad mandates at the national level, such as you see in Japan. Politically that just seems impossible these days. Maybe it will seem more possible in a few years.