Hacker News Reader

Comment by kgeist | Hacker News Reader

I wonder what's their plan when a vehicle gets stuck?

I assume they are planning to have some kind of rescue vehicle. But the problem with this will be the traffic jam the stuck vehicle will cause. Not only will all the vehicles behind it need to stop as well, they also need to be rerouted in order to let the rescue vehicle pass.

The main difference of pumping water or gas through pipes is that the consumer does not care which part of the pumped product he gets. It's all the same. Delivering customer specific goods through pipes requires a very good management of the vehicle flow.

Simply back the other vehicles out, it's not rocket science

At first I was like "good luck getting humans to back out orderly" then I remembered these were autonomous and can just get ordered to go back.

Rocket-powered delivery vehicles would be pretty exciting though.

https://en.wikipedia.org/wiki/Single-line_working

They're likely only running a single tunnel during very early testing in order to keep costs down.

Their idea is basically a cargo-carrying, miniature version of a rubber-tyre metro:

https://en.wikipedia.org/wiki/Rubber-tyred_metro

..which is a century old concept. Thus well-understood, as evidenced by the wikipedia article listing their benefits and downsides versus rail.

Yeah but if a Paris Metro train gets stuck, the tunnels are enormous and your maintenance staff can walk from the nearest station to repair and un-stuck the train.

It's not clear to me that these tunnels will be sufficiently large.

You send a recovery vehicle down the line to fetch it, capable of hooking it, and strong enough to tow/ drag it to an exit point or service "vault" just like "vaults" on underground tunnels for power, data, gas, etc. The tow vehicle basically just needs to weigh enough.

Or you send a human down the line on a platform with wheels. The tunnel would have to be at least as high as those crates and at least twice as wide.

There are far bigger challenges here, like the absolutely massive cost of digging / boring the tunnel and trying to amortize that over what you're delivering. You have to move a fuckton of Big Macs to have this make financial sense, and your competitors include "a guy on an electric bike making tips who uses a few KWhr of electricity over an 8-10 hour shift."

It's really hard to compete against massively subsidized public infrastructure (roads and ICE vehicles) when you're not subsidized at all.

Has any private company in the last decade or so had a successful startup that had to build massive infrastructure?

While I’m no fan of Musk and his management of Twitter has been a shit show, I must begrudgingly admit that he is damn good at building companies that produce real tangible hardware based products that require infrastructure

Ok get this. We make a spider bot that suctions to random cars going in the general direction that we want to deliver the package. If the car turns to go in the wrong direction, the spider bot jumps off and hitches to the next vehicle. Free transit, no infrastructure costs.

Considering this for kicks: People probably wouldn't like their fuel economy being tanked by the extra weight and aerodynamic drag of your spiders. So, as a driver, I might try to make my car slippery or otherwise ungrabbable. Alternatively, I might try to let off the gas so that the other cars end up having to do the work of hauling the spider. Naturally if everyone does this, the whole gang slows to a halt.

You could design all vehicles to be capable of some basic rescue functions.

I would bet the most common situation is the stuck vehicle is just stalled and could be pushed and/or pulled by the car behind or in front of it.

For pulling, you might use an automatic coupler like railways use: https://en.wikipedia.org/wiki/Railway_coupling#Automatic_cou...

> Instead of using suction, which can be unreliable, Pipedream uses electricity to run a small battery-powered autonomous vehicle

This.. has to be a joke? Surely they don't believe a "battery-powered autonomous vehicle" is more reliable than a pneumatic tube network?

Seems like it would be to me. So long as you monitor your battery charge you know the vehicle will get there and it can tolerate many imperfections in the tube. Pneumatics need a tight seal to the tube and thus cannot tolerate imperfections.

> So long as you monitor your battery charge you know the vehicle will get there

ahem, a BEV has __a lot__ more failure points than just "empty battery"...

sending a complex machine down a tube will _always_ carry the risk of the complex machine reading down. pneumatic tube "vehicles" are simple, they can not break down.

I am not an expert, I don't know which one is better or more reliable, but your claim that once you monitor battery charge, everything is hunky dory is wrong.

While what goes into the tub is less complex, the tolerances on the pneumatics are much tighter and so it is less able to handle problems.

I don't know how the factors compare, but tubes are not perfect.

A powerful plunger!

Service pipes?

Reply on news.ycombinator.com