5 ms·

I’m always amazed every time someone tried really hard to invent new technology that’s just a worse train. I love the overall concept, but damn they’re making i

I’m always amazed every time someone tried really hard to invent new technology that’s just a worse train. I love the overall concept, but damn they’re making it way more complicated than it needs to be.

go_elmo 4y ago
have you factored in the cost of human labor in your judgement? I mean swiss-level salaries. Thats 4k+ usd / month for retail workers. Thats what makes autonomous infrastructure cheaper in the long run than having anyone loading anything on an already overcrowded train net.
- scarier 4y ago
  The issue isn’t how automated it is— it’s easy to automate trains and automatically load pallets. Making each car responsible for autonomous navigation in two dimensions and self-loading becomes a much more complicated engineering problem.
- saiya-jin 4y ago
  More like 7-8k/month if you count true employer's expenses on average employee for persons operating & maintaining complex machineries/connected IT systems.
folkrav 4y ago
How is this a worse train?
- scarier 4y ago
  Because it needs far more complicated and expensive cars for no real benefit.
  - ddalex 4y ago
    The benefit is the modularity - trains don't solve the last mile problem, individually router containers on autonomous drives do.
    - andbberger 4y ago
      trains absolutely solve the last mile problem and there is no better example of their modern usage than in switzerland where your ikea would be supplied by a single boxcar once a week
    - scarier 4y ago
      CST isn’t built to solve the last mile problem either—it explicitly shows cargo being delivered to and picked up from central hubs.
    - yakak 4y ago
      This is not a real last mile solution either. We once had the almost the same level of modularity by detaching cars in motion and letting them coast.. Really this is more about not messing with the existing mix of safety and liability defenses used to keep the current situation on the current tracks. These systems have to be sufficiently different because we aren't going to do a flag day to full automation.
kumarsw 4y ago
I recall someone (possibly Adam Something on YouTube complaining about a similar idea) coining the term "Not a Train" to describe concepts such as Hyperloop and this, the idea being that they are (1) futuristic and exiting and (2) because they are different, allow for arguments that the upfront costs will be much lower than trains, whose costs are well studied and known to be high. Usually these systems are more costly and have lower throughput than trains because: 1. Trains use existing, mature technology and there are plenty existing manufacturers 2. Because trains are coupled together into trains, throughput is higher because there is not a need to space self-driving carriages apart 3. Because trains are coupled together, the expensive bits go in the locomotives and the carriage construction is simpler and cheaper
- mschuster91 4y ago
  The problem with trains is that you need large signalling blocks and distances between trains and blocks for safety reasons - a fully loaded freight train can have a braking distance of well over 1.000 meters, and accelerating and decelerating them takes a lot of time and energy. Also, shunting carriages between trains takes up lots of space for the yards, a lot of time to couple/uncouple and verify the correct operation of the brakes. The CST system seems to operate at almost zero distance between carriages, which even with the lower speed of 30 km/h should still have a greater total bandwidth than an equivalent-sized two track railway system. Also, it can operate 24/7 unlike rail systems (because of noise complaints and co-existence with passenger trains), which should further increase the available bandwidth.
  - jcrawfordor 4y ago
    A conventional train and close headways are not an either-or proposition. Freight trains can (and do in places) operate under computerized moving block systems that allow for headway as short as relative braking distance. Automatic train operation with moving blocks is a relatively mature technology available from a couple of major vendors. There are certainly advantages to single-unit cars that avoid the need for shunting, but the maintenance cost on these will be tremendously higher on a per-unit basis. Hard to say that it's a clear win.
  - 4y ago
willyt 4y ago
I think the innovation is the removal of manual handling involved in transferring goods to a train or between trucks at a logistics hub. By making the modules small enough that they will ultimately be able to trundle autonomously at low speed from the loading bay at a small factory to the hub and then travel at relatively high speed to the end destination without requiring sophisticated AI to guess what the humans around it will do. The modules can pop then pop up in dense city centre environments and postal workers can distribute parcels the last few hundred meters or the module can trundle autonomously for the last few hundred meters at very low speeds on battery power. Or likewise trundle around and between manufacturing areas at low speeds transporting parts along a supply chain.
- jsnell 4y ago
  That's not what this is though. These carriages will just be going from delivery hub to delivery hub. The deliveries to hubs will be by trucks. The last mile transport to stores will be by van or bike. And that's a metaphorical last mile; the density of the hubs appears to be pretty low, with e.g. 2-3 hubs in Zürich. The thing they claim makes this work is bundling all deliveries going to the same receiver (across all suppliers) into the same carriage on the sending end, which then makes it easy to optimize that last mile delivery with small vehicles. This doesn't seem very compelling.
bryanlarsen 4y ago
Once upon a time, motors were incredibly expensive. Not just the motors, but the logistics for the motors. The care and feeding of a coal burning steam engine requires lots of space and people with shovels. So factories had a single motor, a giant steam engine that fed a complicated web of pulleys and crankshafts. Electric motors killed those old single motor factories. The same logic results in a single locomotive pulling a giant train. You need a human per locomotive, keeping the messy diesel in one place helps a lot, diesel engines need a lot of maintenance, etc. In 2021, electric motors and batteries are cheap and tiny. Trains on a dedicated track don't need humans. Automated swarms of small motors are the future.
- bombcar 4y ago
  Most US trains have a locomotive engineer and a conductor in the lead cab, and the rest of the power units are remotely controlled and unmanned. That's two employees on the train for 3800 tons of freight. Rail where each car is powered is usually called light rail when used for passengers, but I guess there wouldn't be much stopping light rail being used for freight. The key now isn't the motors but the remote controlling.

go_elmo 4y ago

have you factored in the cost of human labor in your judgement? I mean swiss-level salaries. Thats 4k+ usd / month for retail workers. Thats what makes autonomous infrastructure cheaper in the long run than having anyone loading anything on an already overcrowded train net.

scarier 4y ago
The issue isn’t how automated it is— it’s easy to automate trains and automatically load pallets. Making each car responsible for autonomous navigation in two dimensions and self-loading becomes a much more complicated engineering problem.
saiya-jin 4y ago
More like 7-8k/month if you count true employer's expenses on average employee for persons operating & maintaining complex machineries/connected IT systems.

folkrav 4y ago

How is this a worse train?

scarier 4y ago
Because it needs far more complicated and expensive cars for no real benefit.
- ddalex 4y ago
  The benefit is the modularity - trains don't solve the last mile problem, individually router containers on autonomous drives do.
  - andbberger 4y ago
    trains absolutely solve the last mile problem and there is no better example of their modern usage than in switzerland where your ikea would be supplied by a single boxcar once a week
  - scarier 4y ago
    CST isn’t built to solve the last mile problem either—it explicitly shows cargo being delivered to and picked up from central hubs.
  - yakak 4y ago
    This is not a real last mile solution either. We once had the almost the same level of modularity by detaching cars in motion and letting them coast.. Really this is more about not messing with the existing mix of safety and liability defenses used to keep the current situation on the current tracks. These systems have to be sufficiently different because we aren't going to do a flag day to full automation.

kumarsw 4y ago

I recall someone (possibly Adam Something on YouTube complaining about a similar idea) coining the term "Not a Train" to describe concepts such as Hyperloop and this, the idea being that they are (1) futuristic and exiting and (2) because they are different, allow for arguments that the upfront costs will be much lower than trains, whose costs are well studied and known to be high. Usually these systems are more costly and have lower throughput than trains because: 1. Trains use existing, mature technology and there are plenty existing manufacturers 2. Because trains are coupled together into trains, throughput is higher because there is not a need to space self-driving carriages apart 3. Because trains are coupled together, the expensive bits go in the locomotives and the carriage construction is simpler and cheaper

mschuster91 4y ago
The problem with trains is that you need large signalling blocks and distances between trains and blocks for safety reasons - a fully loaded freight train can have a braking distance of well over 1.000 meters, and accelerating and decelerating them takes a lot of time and energy. Also, shunting carriages between trains takes up lots of space for the yards, a lot of time to couple/uncouple and verify the correct operation of the brakes. The CST system seems to operate at almost zero distance between carriages, which even with the lower speed of 30 km/h should still have a greater total bandwidth than an equivalent-sized two track railway system. Also, it can operate 24/7 unlike rail systems (because of noise complaints and co-existence with passenger trains), which should further increase the available bandwidth.
- jcrawfordor 4y ago
  A conventional train and close headways are not an either-or proposition. Freight trains can (and do in places) operate under computerized moving block systems that allow for headway as short as relative braking distance. Automatic train operation with moving blocks is a relatively mature technology available from a couple of major vendors. There are certainly advantages to single-unit cars that avoid the need for shunting, but the maintenance cost on these will be tremendously higher on a per-unit basis. Hard to say that it's a clear win.
- 4y ago

willyt 4y ago

I think the innovation is the removal of manual handling involved in transferring goods to a train or between trucks at a logistics hub. By making the modules small enough that they will ultimately be able to trundle autonomously at low speed from the loading bay at a small factory to the hub and then travel at relatively high speed to the end destination without requiring sophisticated AI to guess what the humans around it will do. The modules can pop then pop up in dense city centre environments and postal workers can distribute parcels the last few hundred meters or the module can trundle autonomously for the last few hundred meters at very low speeds on battery power. Or likewise trundle around and between manufacturing areas at low speeds transporting parts along a supply chain.

jsnell 4y ago
That's not what this is though. These carriages will just be going from delivery hub to delivery hub. The deliveries to hubs will be by trucks. The last mile transport to stores will be by van or bike. And that's a metaphorical last mile; the density of the hubs appears to be pretty low, with e.g. 2-3 hubs in Zürich. The thing they claim makes this work is bundling all deliveries going to the same receiver (across all suppliers) into the same carriage on the sending end, which then makes it easy to optimize that last mile delivery with small vehicles. This doesn't seem very compelling.

bryanlarsen 4y ago

Once upon a time, motors were incredibly expensive. Not just the motors, but the logistics for the motors. The care and feeding of a coal burning steam engine requires lots of space and people with shovels. So factories had a single motor, a giant steam engine that fed a complicated web of pulleys and crankshafts. Electric motors killed those old single motor factories. The same logic results in a single locomotive pulling a giant train. You need a human per locomotive, keeping the messy diesel in one place helps a lot, diesel engines need a lot of maintenance, etc. In 2021, electric motors and batteries are cheap and tiny. Trains on a dedicated track don't need humans. Automated swarms of small motors are the future.

bombcar 4y ago
Most US trains have a locomotive engineer and a conductor in the lead cab, and the rest of the power units are remotely controlled and unmanned. That's two employees on the train for 3800 tons of freight. Rail where each car is powered is usually called light rail when used for passengers, but I guess there wouldn't be much stopping light rail being used for freight. The key now isn't the motors but the remote controlling.