10 ms·

Hmm I don't know. It's been 10 years since I've been hearing that self driving cars are right around the horizon. And we now know we're nowhere close to it. If

by Jagat 7y ago

Hmm I don't know. It's been 10 years since I've been hearing that self driving cars are right around the horizon. And we now know we're nowhere close to it. If this recaptcha helps google, or for that matter, any company, accelerate their self driving capability, I'd fully support it.

garaetjjte 7y ago
https://xkcd.com/1897/ https://xkcd.com/1897/
djsumdog 7y ago
Billions of tax dollars and breaks are going to self-driving research, which is fucking insane. America needs to build back it's public rail infrastructure. Passenger rail is a solved problem, and the US use to have more passenger rail that Europe. We have one of the best freight rail systems in the world, but we have high speed rail in New England and Florida. That's it. Maybe California if they can get their shit together. I wrote an article about this before, and talk about how even if you had a four lane high way with nothing but self driving cars, all filled and traveling at 120kph bumper to bumper, you wouldn't even come close to approaching 10% of the capacity of a single track metro light rail line running with 2 minute headway (during rush hour headways can be less than 2min, and with automated systems like those in Singapore and the new ones coming to London's underground--DLR is already automated--you can get headways of less than a minute or even 30 seconds) https://penguindreams.org/blog/self-driving-cars-will-not-solve-the-transportation-problem/ https://penguindreams.org/blog/self-driving-cars-will-not-so... Sure self driving cars could help in Europe with the last leg where they have real transport infrastructure, but the US is so far behind that successful self driving cars would just add to grid-lock.
- toasterlovin 7y ago
  Some thoughts: I think rail adoption becomes a lot more feasible if people can summon self driving cars on demand. The problem with rail is that it's hard for people who are currently driving to get behind because its usefulness is dependent on how built out the network is and building out rail networks is slow and expensive. And while the network has low coverage, getting to and from rail stations is a pain in the ass. But if you can just hail an Uber or Lyft, but the Uber or Lyft is really cheap because it doesn't require a human driver, then getting to and from the rail station becomes a non-issue. And I think that can turn the all or nothing paradigm of rail as it currently exists into something where incremental progress is just as useful whether you're starting from 0% existing coverage or 80% existing coverage.
  - majormajor 7y ago
    > But if you can just hail an Uber or Lyft, but the Uber or Lyft is really cheap because it doesn't require a human driver, then getting to and from the rail station becomes a non-issue. We're a long ways away from self driving cars being really cheap. That self driving Uber network would need a lot of capital investment that they initially were able to avoid because they piggybacked on drivers already having to own cars anyway.
    - toasterlovin 7y ago
      I agree that A) the technology is not there, and B) it would require a lot of capital. I still think it's the most realistic approach to passenger rail utilization in most of the U.S. It will take decades to build out high coverage passenger rail in the U.S. That's a lot of time for technology to mature and for capital to be deployed.
- manigandham 7y ago
  This discussion has been well hashed before, but comes down to the US being very big. People are far more spread out here than anywhere in Europe or Asia. This also makes last-mile transport even more critical if you aren't close to the station. The distances also drive up the cost and waits between trains, further reducing ridership. High-speed mid/long distance passenger rail just isn't viable given our population densities, and we already have metros/subways/light-rail in core metropolitan areas. Without major construction with hundreds of new lines and a shift in city planning, it's unlikely to ever change in the US.
  - djsumdog 7y ago
    That is the worst possible argument. Australia has a fraction of our population and every one of their capitals (except for Darwin and Hobart) has a very good rail system. Sure they don't have high speed, but they could easily saturate a Sydney to Melbourne high speed route (if Melbourne didn't waste a few billion on their ticketing system). Russia has high speed rail and it's less dense than the United States. China created their high speed system in less than a decade. This argument comes up all the time and it's so poor. The United States use to have more passenger rail than Europe does now! You build light city rail and immediately, new housing and commercial stuff pops up around it. It can potentially reduce drunk driving as well. The US/density argument is really tired and just doesn't hold up when you really look at it.
    - andrewshadura 7y ago
      Correction: Russia doesn't really have high-speed trains, except one tiny line which doesn't count if you consider the size of the country.
    - robbiep 7y ago
      It’s quite generous to say that Australian capitals have very good rail systems. They exist, sure. Patronage (especially in sydney) has been growing at a huge level (partially due to the traffic nightmare, a side effect of the geography). But it’s a long way from good or optimal. But yes, if we built high speed rail linking Melbourne - Canberra - sydney - Brisbane we would have one hell of an east coast system
- cntrlaltdlt 7y ago
  I am certainly on your side when it comes to mass rail transit being a really good solution for inter-city transportation. Especially now in the age where intra-city transit is so easy with uber/lyft. The problem as it stands is land easements. Acquiring land in the US for larger social benefit at the expense of the land owner is generally A BIG FUCKIN NO NO. This premise has essentially locked many well planned rail projects from the get go. This just isn't a problem for the promise of self-driving car technology. Further more the opportunity cost argument is really weak. Al la "had we not spent so much money trying invent self driving cars we could have improved our rail system X times over ect." Because we have developed so much other technology out of that quest. I dream of a day where there is a high speed rail system that connects the entire west coast. It would be a GIGANTIC boon for trade and commerce IMO.
wolco 7y ago
Solving self driving will go hand in hand with solving flying cars. When cars can fly self driving is trivial and lanes above can be controlled. The streets can go back to bikes and walking and for older cars.
- lurker458 7y ago
  flying requires an order of magnitude more energy compared to driving. It won't be be a feasible method of mass transportation anytime soon.
  - wolco 7y ago
    That's how long it will take for mass adoption of self driving cars.
cptnapalm 7y ago
Self-driving cars were right around the corner in the 1980s, too. There was an AI conference shortly before the last AI Winter where this was remarked upon. I'm kicking myself for not saving the transcript; self-driving cars, worries of regular people losing their jobs, clueless reporters grossly inflating regular people's expectations, etc.
cptnapalm 7y ago
Found it: https://pdfs.semanticscholar.org/2a85/7e042331c53b856b3ad5018454f72d798790.pdf https://pdfs.semanticscholar.org/2a85/7e042331c53b856b3ad501...

by Jagat 7y ago

garaetjjte 7y ago

https://xkcd.com/1897/ https://xkcd.com/1897/

djsumdog 7y ago

Billions of tax dollars and breaks are going to self-driving research, which is fucking insane. America needs to build back it's public rail infrastructure. Passenger rail is a solved problem, and the US use to have more passenger rail that Europe. We have one of the best freight rail systems in the world, but we have high speed rail in New England and Florida. That's it. Maybe California if they can get their shit together. I wrote an article about this before, and talk about how even if you had a four lane high way with nothing but self driving cars, all filled and traveling at 120kph bumper to bumper, you wouldn't even come close to approaching 10% of the capacity of a single track metro light rail line running with 2 minute headway (during rush hour headways can be less than 2min, and with automated systems like those in Singapore and the new ones coming to London's underground--DLR is already automated--you can get headways of less than a minute or even 30 seconds) https://penguindreams.org/blog/self-driving-cars-will-not-solve-the-transportation-problem/ https://penguindreams.org/blog/self-driving-cars-will-not-so... Sure self driving cars could help in Europe with the last leg where they have real transport infrastructure, but the US is so far behind that successful self driving cars would just add to grid-lock.

toasterlovin 7y ago
Some thoughts: I think rail adoption becomes a lot more feasible if people can summon self driving cars on demand. The problem with rail is that it's hard for people who are currently driving to get behind because its usefulness is dependent on how built out the network is and building out rail networks is slow and expensive. And while the network has low coverage, getting to and from rail stations is a pain in the ass. But if you can just hail an Uber or Lyft, but the Uber or Lyft is really cheap because it doesn't require a human driver, then getting to and from the rail station becomes a non-issue. And I think that can turn the all or nothing paradigm of rail as it currently exists into something where incremental progress is just as useful whether you're starting from 0% existing coverage or 80% existing coverage.
- majormajor 7y ago
  > But if you can just hail an Uber or Lyft, but the Uber or Lyft is really cheap because it doesn't require a human driver, then getting to and from the rail station becomes a non-issue. We're a long ways away from self driving cars being really cheap. That self driving Uber network would need a lot of capital investment that they initially were able to avoid because they piggybacked on drivers already having to own cars anyway.
  - toasterlovin 7y ago
    I agree that A) the technology is not there, and B) it would require a lot of capital. I still think it's the most realistic approach to passenger rail utilization in most of the U.S. It will take decades to build out high coverage passenger rail in the U.S. That's a lot of time for technology to mature and for capital to be deployed.
manigandham 7y ago
This discussion has been well hashed before, but comes down to the US being very big. People are far more spread out here than anywhere in Europe or Asia. This also makes last-mile transport even more critical if you aren't close to the station. The distances also drive up the cost and waits between trains, further reducing ridership. High-speed mid/long distance passenger rail just isn't viable given our population densities, and we already have metros/subways/light-rail in core metropolitan areas. Without major construction with hundreds of new lines and a shift in city planning, it's unlikely to ever change in the US.
- djsumdog 7y ago
  That is the worst possible argument. Australia has a fraction of our population and every one of their capitals (except for Darwin and Hobart) has a very good rail system. Sure they don't have high speed, but they could easily saturate a Sydney to Melbourne high speed route (if Melbourne didn't waste a few billion on their ticketing system). Russia has high speed rail and it's less dense than the United States. China created their high speed system in less than a decade. This argument comes up all the time and it's so poor. The United States use to have more passenger rail than Europe does now! You build light city rail and immediately, new housing and commercial stuff pops up around it. It can potentially reduce drunk driving as well. The US/density argument is really tired and just doesn't hold up when you really look at it.
  - andrewshadura 7y ago
    Correction: Russia doesn't really have high-speed trains, except one tiny line which doesn't count if you consider the size of the country.
  - robbiep 7y ago
    It’s quite generous to say that Australian capitals have very good rail systems. They exist, sure. Patronage (especially in sydney) has been growing at a huge level (partially due to the traffic nightmare, a side effect of the geography). But it’s a long way from good or optimal. But yes, if we built high speed rail linking Melbourne - Canberra - sydney - Brisbane we would have one hell of an east coast system
cntrlaltdlt 7y ago
I am certainly on your side when it comes to mass rail transit being a really good solution for inter-city transportation. Especially now in the age where intra-city transit is so easy with uber/lyft. The problem as it stands is land easements. Acquiring land in the US for larger social benefit at the expense of the land owner is generally A BIG FUCKIN NO NO. This premise has essentially locked many well planned rail projects from the get go. This just isn't a problem for the promise of self-driving car technology. Further more the opportunity cost argument is really weak. Al la "had we not spent so much money trying invent self driving cars we could have improved our rail system X times over ect." Because we have developed so much other technology out of that quest. I dream of a day where there is a high speed rail system that connects the entire west coast. It would be a GIGANTIC boon for trade and commerce IMO.

wolco 7y ago

Solving self driving will go hand in hand with solving flying cars. When cars can fly self driving is trivial and lanes above can be controlled. The streets can go back to bikes and walking and for older cars.

lurker458 7y ago
flying requires an order of magnitude more energy compared to driving. It won't be be a feasible method of mass transportation anytime soon.
- wolco 7y ago
  That's how long it will take for mass adoption of self driving cars.

cptnapalm 7y ago

Self-driving cars were right around the corner in the 1980s, too. There was an AI conference shortly before the last AI Winter where this was remarked upon. I'm kicking myself for not saving the transcript; self-driving cars, worries of regular people losing their jobs, clueless reporters grossly inflating regular people's expectations, etc.

cptnapalm 7y ago

Found it: https://pdfs.semanticscholar.org/2a85/7e042331c53b856b3ad5018454f72d798790.pdf https://pdfs.semanticscholar.org/2a85/7e042331c53b856b3ad501...