6 ms·

Nah. It's long-haul trucking. There's a much easier on-ramp to profitability because you don't have to solve the whole problem at once. You start by just autom

Nah. It's long-haul trucking. There's a much easier on-ramp to profitability because you don't have to solve the whole problem at once. You start by just automating the highway driving, which is a much easier problem than city driving. You might still need the human operator to deal with the off-highway driving at the beginning and end of each journey, but in the middle they can just hang out in the cab and work on their PhD dissertation or whatever. At that point your business has become quite a bit more efficient, because you've gotten rid of the need for your trucks to be stationary for 10 out of every 24 hours. So that's already a big win for tackling a relatively small piece of the problem. You can keep chipping away from there, and each increment will net even more returns.

madeofpalk 10y ago
I thought long-haul trucking was a given. I'm still believing that taxi/uber service will be the first consumer facing application.
motoboi 10y ago
They way I see it, companies just need a place outside the city where drivers can enter the truck and do the last part (or leave the truck and get the next one). This may work similar to ports, where a pilot come to the ship to navigate the port area.
- jamiek88 10y ago
  Or even better they park and the trailers are then hooked up to human driven, electric tugs that aren't realistic for cross country highway travel (range/density of battery) but are for under 50 mile trips into the city.
- Spooky23 10y ago
  Robot trucks turn the economics of trucking company into something more like railroads, which having been eating the lunch of trucking for awhile now.
  - ghaff 10y ago
    >railroads, which having been eating the lunch of trucking for awhile now Right. Which means that a lot of trucking isn't especially long haul already which in turn means that taking people out of only the highway portion isn't necessarily the big win that a lot of people assume it could be. If you still need truckers around the endpoints, it doesn't necessarily make sense to take them out of the middle (given other advantages to having a human operator). Of course, good assistive systems can still be used to improve safety.
    - Spooky23 10y ago
      That's why I worry about this area of technology, especially in the context of the history of AI in the past. The hype factor is extreme, but there's very little domain expertise at play. I have a relative who is a trucking regulation consultant. It's not some trivial business... there are a lot of moving parts.
IanCal 10y ago
> You start by just automating the highway driving, which is a much easier problem than city driving. You might still need the human operator to deal with the off-highway driving at the beginning and end of each journey, but in the middle they can just hang out in the cab and work on their PhD dissertation or whatever. At that point your business has become quite a bit more efficient, because you've gotten rid of the need for your trucks to be stationary for 10 out of every 24 hours. Can't this already be handled in a pony-express fashion? Hand off the trucks between people? Honest question, I assumed that this kind of thing might already be setup. Your solution also requires that a person sit in a single box for 24 hours, I can see that taking a toll, and the start and end of the journey must be timed that a person can sleep at the right time before waking up ready to drive again as the truck needs to come off the highway. I guess my key thoughts are that this doesn't actually lower the human cost (employed for the full time still), and adds on a few extra issues dealing with people. What you gain is a faster overall transport. However, I can see the added automation cost being a lower part of the overall vehicle bill, and there are enough large trucking companies to mean the costs can be spread out over the business.

madeofpalk 10y ago

I thought long-haul trucking was a given. I'm still believing that taxi/uber service will be the first consumer facing application.

motoboi 10y ago

They way I see it, companies just need a place outside the city where drivers can enter the truck and do the last part (or leave the truck and get the next one). This may work similar to ports, where a pilot come to the ship to navigate the port area.

jamiek88 10y ago
Or even better they park and the trailers are then hooked up to human driven, electric tugs that aren't realistic for cross country highway travel (range/density of battery) but are for under 50 mile trips into the city.
Spooky23 10y ago
Robot trucks turn the economics of trucking company into something more like railroads, which having been eating the lunch of trucking for awhile now.
- ghaff 10y ago
  >railroads, which having been eating the lunch of trucking for awhile now Right. Which means that a lot of trucking isn't especially long haul already which in turn means that taking people out of only the highway portion isn't necessarily the big win that a lot of people assume it could be. If you still need truckers around the endpoints, it doesn't necessarily make sense to take them out of the middle (given other advantages to having a human operator). Of course, good assistive systems can still be used to improve safety.
  - Spooky23 10y ago
    That's why I worry about this area of technology, especially in the context of the history of AI in the past. The hype factor is extreme, but there's very little domain expertise at play. I have a relative who is a trucking regulation consultant. It's not some trivial business... there are a lot of moving parts.

IanCal 10y ago

> You start by just automating the highway driving, which is a much easier problem than city driving. You might still need the human operator to deal with the off-highway driving at the beginning and end of each journey, but in the middle they can just hang out in the cab and work on their PhD dissertation or whatever. At that point your business has become quite a bit more efficient, because you've gotten rid of the need for your trucks to be stationary for 10 out of every 24 hours. Can't this already be handled in a pony-express fashion? Hand off the trucks between people? Honest question, I assumed that this kind of thing might already be setup. Your solution also requires that a person sit in a single box for 24 hours, I can see that taking a toll, and the start and end of the journey must be timed that a person can sleep at the right time before waking up ready to drive again as the truck needs to come off the highway. I guess my key thoughts are that this doesn't actually lower the human cost (employed for the full time still), and adds on a few extra issues dealing with people. What you gain is a faster overall transport. However, I can see the added automation cost being a lower part of the overall vehicle bill, and there are enough large trucking companies to mean the costs can be spread out over the business.