Hacker News Reader

Comment by schiffern | Hacker News Reader

  >rail

Wrong niche. This is in the same market as conventional air freight.

View on HN

How much air freight really needs to be transferred by air, though?

Air freight makes sense when you need to get something across an ocean in less than a month, or across the continent in less than a day. If the US were to invest in running the railroads efficiently, getting a package from NYC to LA in less than 72 hours should be possible - that's fast enough for regular Amazon shipping. If you electrify the railroads you could avoid an awful lot of CO2 emissions.

There are currently already container trains running from China to Europe (or at least there used to, before Russia started a war). It's a viable midpoint between ocean freight and air freight.

> If the US were to invest in running the railroads efficiently, getting a package from NYC to LA in less than 72 hours should be possible

Still need air for every other pair of destinations. And in the meantime, I’d be shocked if that freight leg had enough volume to justify that expense over a couple more planes.

US freight rail companies have already invested heavily in efficiency and are now pretty much the most efficient in the world. What you seem to be asking for is a reduction in latency. You can't have both high efficiency and low latency: those metrics are in direct conflict with each other.

Trucks with team driving can already run coast to coast in about 3 days. This works fine for somewhat time sensitive loads.

I think freight trains in Europe are much faster than in the US, but also much shorter (though there are plans to make them a little longer) and lighter (no double-stacked containers). So they are more costly but you get stuff from A to B faster.

Electrified locomotives are better than diesel electric (e.g. an electric locomotive has 2x - 3x the power, ~6MW-9MW compared to ~3MW, so you need less locomotives, less moving parts means less maintenance etc.), but while it works in dense Europe, it's too costly to build and maintain cross US.

it's probably too expensive for some of the country (mainly the great plains etc), but on the west coast and east of the Mississippi there's definitely the density to support it.

72 hours is too long. You need 48 hours or less for e-commerce or you’ll lose business to Amazon. You could go multiple warehouses but that’s a huge expense with splitting inventory especially if you have lots of low volume skus.

Air freight is phenomenally inefficient compared to rail.

Most of the time but some remote locations it is not nor does it require rail volume.

Rail is less efficient than water transport most of the time but it does not always apply to the local situation.

Pareto principle means you should prioritize the 80%+ use case. Most of the time is what you want to spend all your time/money optimizing while making also the other use cases possible.

I think this is a terrible way to apply the Pareto principle.

80% of the time a hammer is a better tool than a screwdriver does not mean you should not use screwdrivers and just get better with a hammer.

Different tools for the job is more effective than one universal tool for everything

Rail is phenomenally inefficient crossing oceans compared to air freight

Hmm, if we ever built the Alaska-Siberia overland railway connection (bridges + tunnels) it would become a possibility to make efficient from what is now an impossibility.

It’s doable. Tunnels and bridges necessary would not be beyond what we have built elsewhere with current technology.

The weather may be challenging in Winter though.

How is building a rail system for transporting freight to and from the U.S. that runs through Russian territory a good/feasible idea?

Even if Russia agreed, Putin or whatever strong-man dictator they have next could suddenly decide it was a bad idea, and to block U.S. freight. Then we have no rail transport AND a grounded air fleet that would take time to get operational at the same output it is today.

I didn't mean tomorrow ("if we ever.."), or the day after. I meant one day when the US and Russia get along as well as China and the US do. It's not like we'll be at not-war war long into the future. Building the bridges and tunnels would take probably a decade or two, the way we build things. But if we did, then much of China's trade that goes over airfreight could come over via rail... one day.

When goods stop crossing borders, soldiers start crossing borders. The easiest way NOT to go to war is to have open trade.

Best way to avoid war is to have the border be open ocean.

Having a railroad bridge with russia is inviting problems

That's a nice little platitude but it's not supported by the historical record.

Eh, the USA’s history seems to confirm it. I’m not really sure about others, but I’d be surprised if countries that are dependent upon one another were at war with one another.

Where is Mr Wilford when you need him?

Sorry couldn’t resist the Snowpiercer Reference - where they connect alaska to russia with a long ass bridge.

To be fair it does look like a “technically possible but ruinously expensive” kinda project.

If there was actually demand for faster high-volume shipping between East Asia and North America then logistics companies would just buy faster ships. The technology for that already exists and is far cheaper than building a railroad tunnel across the Bering Strait. Shipyards are capable of building freighters with more hydrodynamic hulls and powerful turbine engines. So far no one is really asking for those, which indicates the market doesn't exist.

it exists in the form of air freight, which is stupidly more expensive, but also stupidly faster. so there is a market, just apparently not one in-between.

That’s several thousand miles of additional journey on your way to Shenzen or wherever? Massive amounts of track, tunnel and bridge that has to be maintained in extreme conditions and across multiple nations at odds with each other?

Or just put it on a boat?

I don't think it would be as problematic as you suggest. The Trans-Siberian railway already operates at similar distances. Also China is building their belt-and-road network. We could get goods from China, Japan, Korea and Russia all on this rail system.

The one sticky point might be gauge. We'd want something all agree upon. We'd want to avoid variable track gauge bogies.

Clearly, we need more airship transport.

Travel time of 8 hours vs 3 weeks... most things being transported by plane are ones where time is of the essence.

There’s such a huge time gap between our current rail and air that many things that are somewhat time sensitive end up flying which would use a reasonably quick rail network. NYC to LA is well under 2,800 miles so getting just about anywhere in the continental US on rail in under 48 hours is a completely reasonable standard.

Yet, people moving from Texas to Florida don’t just park a U-Haul on a flatbed and fly, because it’s not just slow but also unpredictable. It can take weeks or even months in some cases because the network is optimized for coal and wheat etc which don’t care about delays just cost.

You think a completely reasonable standard is for a train to average 58 mph all the way across the united states without stopping?

What about all the slowdowns due to extremely rough terrain? The fact people have to be swapped out? The fact that not all cars are going all the way across the country and also have to be swapped out?

58mph average is insane.

Train’s don’t need to slow down for rough terrain because they can’t handle rough terrain and either route around it or tunnel through it. What they do need to slow down for is the rail network itself.

Poor track conditions, missing block signal systems, and Trains without an automatic cab signal, automatic train stop or automatic train control system "may not exceed 79 mph."

Freight trains really could travel a mostly Class 6 network at 120MPH in the US, we have regulations all the way to class 9. They don’t because that’s not rail is optimized for.

So as you said it's not actually reasonable to think a train is going to get across the US in 48 hours.

Trains have to slow down a lot when going through rough terrain because there are curves involved. You can't go 80mph.

Reply on news.ycombinator.com