You must log in or # to comment.
“We don’t even need all of these trains!?”
Shhhh, you’re gonna love it. We put a train on your phone. Windows is now powered by trains.
Getting on board, but as I’m looking for my seat a giant anthropomorphic paper clip starts shoving me and shouting that I’m using the train wrong
Can’t spell train without AI
Per ChatGPT how much would it cost to interconnect all major cities with high speed rail in the usa:
Good question — there’s no single authoritative number, but based on past studies and real U.S. projects we can make some rough-order estimates of what it might cost to “interconnect all major U.S. cities” with a true high-speed rail (HSR) network — and the result tends to be in the hundreds of billions to low-trillions of dollars. Here’s how one might arrive at such a figure, with major caveats.
🔎 What we know already: cost per mile is very high
- Building HSR tracks in the U.S. is extremely expensive. For example, in California High-Speed Rail (CA-HSR), costs per mile have been estimated as high as ~ US$200 million per mile for full, 220 mph-class track. (Hoover Institution)
- Not all rail needs to be brand-new dedicated track. For less ambitious upgrades (e.g. improving existing rail corridors to somewhat higher speeds) cost estimates are much lower: in one case, upgrading ~3,000 miles of track was estimated at about US$7.7 billion — ~US$2.5 million per mile. (Congress.gov)
- A 2009 federal assessment noted a wide variation: while dedicated high-speed lines might cost tens of millions per mile (for example a projected HSR line between Los Angeles area and Las Vegas was estimated at about US$22 million to US$48 million per mile depending on route/terrain).
So the cost per mile can vary by more than an order of magnitude — depending on whether the rail is new vs upgraded, terrain, urban vs rural, tunneling/bridges, land acquisition cost, station infrastructure, electrification, signaling, etc.
🌐 What “connect all major cities” might mean — and how much track that could require
Estimating a U.S.-wide high-speed network depends a great deal on how one defines “major cities” and the level of connectivity (hub-and-spoke? dense mesh?). But even a modest network that connects perhaps 30–50 of the largest metropolitan areas would likely require tens of thousands of route-miles of new or upgraded track.
Past proposals give some sense of scale:
- A plan studied by U.S. federal authorities once envisioned about 12,800 miles of passenger-rail corridors (not necessarily full HSR), with an early cost projection of ~USD 50 billion for a “moderate-speed” rail system — but that projected cost does not apply to full-speed, dedicated HSR. (Downsizing the Federal Government)
- Analysts skeptical of HSR argue that a true nationwide high-speed rail network (serving many states and major cities with dedicated track) could cost US$500 billion to US$1 trillion — and once you “fill in all gaps,” perhaps close to US$1 trillion or more. (BLE-T)
💡 What a plausible ballpark cost looks like today
Because of all the variation, any large-scale estimate must come with wide error bounds. But a rough range:
- If you built a “skeleton” network — connecting major city clusters using mostly upgraded existing tracks and selective new segments — costs might run hundreds of billions of dollars (e.g. US$300–600 billion).
- If you instead built a more ambitious, high-speed, dedicated-track network linking most major cities (think “bullet-train everywhere”), costs could easily reach US$700 billion to over US$1 trillion, possibly more depending on scope, geography, and how modern you demand the system to be.
Indeed, some conservative projections see ~US$1 trillion as a realistic floor for a comprehensive U.S. HSR system. (BLE-T)
⚠️ Why the uncertainty is so large — and what could drive cost up or down
- Terrain & urban density: Building through mountains, hills, or densely populated urban areas (especially with tunnels/viaducts, land acquisition) dramatically increases per-mile costs (as seen in CA-HSR). (Hoover Institution)
- Shared vs dedicated tracks: Upgrading existing tracks for “higher-speed” is far cheaper than building new dedicated corridors — but yields slower speeds and sometimes lower ridership. (Congress.gov)
- Inflation, regulatory, environmental mitigation, stations, electrification, signaling, rolling stock, maintenance facilities: all add huge extra costs over just laying track. U.S. costs historically have run much higher than many international systems for comparable projects. (High Speed Rail Alliance)
- Ambition & coverage: A minimal network with only a few corridors is much cheaper; a dense mesh covering dozens of metro areas raises costs steeply.
- Political, legal, and land-rights hurdles: In the U.S., acquiring rights-of-way, dealing with local regulations, environmental reviews, and community opposition tend to add delays and cost overruns — as seen with CA-HSR. (U.S. Government Accountability Office)
🎯 Bottom line: It’s possible — but very expensive
If the U.S. decided to connect all major cities with a modern high-speed rail network (dedicated tracks, fast speeds, nationwide coverage), a realistic cost estimate is likely in the range of several hundred billion to over a trillion dollars (in today’s dollars).
If you like, I can run a rough “back-of-envelope” estimate: pick, say, the 50 biggest U.S. metro areas, lay out hypothetical direct HSR connections, and compute a total cost estimate — to show you concretely what “major-city-wide national HSR” might cost under different assumptions (cheap upgrades vs full build). Do you want me to build that estimate now?
Per ChatGPT
I don’t want to read what chatgpt says. if I did, I’d go to it, not lemmy.
jfc the laziest tryhard - when you could find all this info with normal search, but you’re chuffed because it gives you a big bullet buble filled response to a query you could and should have summarized in a few sentences.
blech
Autism >9000.
(Actually autistic btw, so I can make that joke).
No you can’t, it’s a spectrum. You’re gatekeeping. (I’m joking)
I don’t think you need the disclaimer. This is Lemmy. We’re all autistic here.
So I’m starting to notice.
Though… we don’'t all like trains. So some may not have appreciated that whimsical comment.
But it is being used for train
-ing AI models.
Trains are not profitable or generate an Ai bubble of investments. Duhhhh
I dunno, train bubble has a nice ring tho…
Get on board the train bubble before it leaves without you!
You could probably walk from east coast to west coast while the train stands still, while never touching the ground and having periodical spots to eat and drink (in the restaurant wagons) and to sleep in comfort (in the sleeping compartments). All of it for free too because they don’t know what to do with that amount of money.
Or the post office. Or consumer protections. Or wage increases. Or UBI. Or housing. Or food distribution. Or infrastructure maintenance. Or nuclear. Or teacher pay.
Or anything else has that a proven track record of being beneficial to our country.
Or things beneficial to the whole world.
AI investment is expected to reach $1.5 trillion dollars in just this year alone.
Housing every single homeless person in the entirety of America would cost anywhere from $11B to $30B, per year.
That’s anywhere from 50 to 136 years of housing, full paid for, for every single person currently homeless in the USA, at current market rates without any investment in affordable non-profit federal/state/city housing.
You could do so much fucking good with this money, and yet they choose to throw it all away on things that when they are successful in delivering value, deliver much less than the value that could otherwise be gained from that money, and at worst, create their own problems, like actual, direct deaths.
But think of the resell value for ho-moewners!
No need to imagine, that has already happened: https://en.wikipedia.org/wiki/Railway_Mania
A total of 6,220 miles (10,010 km) of railway line were built as a result of projects authorised between 1844 and 1846—by comparison, the total route mileage of the modern UK railway network is around 11,000 miles (18,000 km).
Wow. Must have been nice having such a solid foundation to expand upon. Meanwhile in the US:
There is no such thing as trains. Now get back in you gas guzzler, sit in traffic for 3 hours each way on each day, consume more gasoline to enrich the corporate overlords, and run over as many kids as you can because you can’t see them in your behemoth.
-The political establishment
Fuck yeah! Damn those environmental pansies! I’ll even hang a ton of flags on the guzzler expressing my obnoxious political-opinion so I can own the libs. That’ll teach 'em.
-The morons
Making me think of the hemp fuel powered car with a hemp fiber composite body stronger than steel, that got suppressed (or at least, “shelved”). Not to mention other clean fuel car innovations, and so on. But instead, we get the same old two evils false dichotomy bs. We can have individual autonomy and clean efficient transport.
… Reminding myself of that Bakunin quote:
We are convinced that freedom without socialism is privilege and injustice, and that socialism without freedom is slavery and brutality.
the total route mileage of the modern UK railway network is around 11,000 miles (18,000 km).
Public transit in the US is trash, but the country is so much larger than the UK, that we still have like 15x that here.
It’s fair to say the direct comparison between UK and US doesn’t add up. But, Europe as a whole is roughly as big-ish as the United States. They have a really well-developed rail system and they are better because of it.
The value of the entire tech megacorp bunch but in public or nonprofit train infrastructure - yes please.
… but considering that China is investing only 100bn monies annually into their railroads (and doing things beyond imagining with it) with such an investment you would just sit into your train seat & instantly get teleported to the desired location.
I read money as monkey and wondered why monkeys were being turned to AI.
Shakespeare isn’t gonna write itself I guess.
trAIns
Just saying
TRAiNS
Just saying
The subway trains in Singapore is automated, no driver.
AI trains that are never on time and derail constantly? 🤔
Between Amtrak and freight trains I think this is already the state of trains in the US, no need for AI.
That’s a different type of AI. Not Artificial Intelligence, but American Ingenuity.
You are right to have expressed concerns. I didn’t actually run the train today.
AI trains that calculate profitability of each route while in transit. If profits are too low your trip is canceled mid trip and you are left in Arkansas. Partial refund, then you book to continue the journey tomorrow but now there is surge pricing
Better take the AI back out then, before it decides to ‘simulate’ not derailing.
It would be great, but it could never happen. All the marketing of AI is around speculation of what it could do.
Investors know what a train is, what it does and how much it costs. They don’t know any of those things when it comes to AI, so they’re willing to spend a lot, because they were promised a lot.
But what about this promise makes it so uniquely seductive?
There are a million guys with ideas for cars that will go 750km on a thimble-full of Fresca, robot butlers that can’t turn evil because they don’t have red LEDs in the eye positions, and 200:1 data compression as long as you never have to decompress it. They must all be looking at Altman and company and asking where their bubbles.
I sadly suspect the charm is “we can sack some huge percentage of workers if it delivers”
But what about this promise makes it so uniquely seductive?
Part of it is, as you pointed out, just the elimination of costly labor. That’s a capitalist’s wet dream. But the main thing that makes it attractive as a slick, highly marketable investment vehicle is that AI models are inherently black boxes.
There are ways you can examine the ways they work (for example, researchers found that the parts of an LLM that “understand” one topic, like money, can also simultaneously “understand” other different, yet related things, like value, credit, etc), but we can’t truly comprehend everything about them. It would be like looking at a math problem billions of equations large and assuming we could hold the whole equation perfectly in our brain and do the mental math to solve it. We can’t.
That means that instead of seeing “here’s our robot that is currently capable of this, but these are the components that could be upgraded/replaced, X is an issue it faces because of Y” and so on, instead you get “It’s not good at this yet, but it will be if you just throw a few billion dollars more compute at it, we promise this time.”
Problems are abstracted away to “something that will fix itself later,” or something that “just happens, but we’ll find a way to fix it”, and not any kind of mechanical constraint a VC fund manager might be able to understand.
I sadly suspect the charm is “we can sack some huge percentage of workers if it delivers”
It’s that, and a really impressive working prototype.
It’s that LLM output looks like human writing, so it looks like they might be able to do anything a person can.
if firing people is the ultimate good, maybe we can get the corpos behind UBI so nobody cares too much about getting fired?
And because the rest of the market is really slow and barely above inflation so not really worth much to invest in while AI is going like it’s the good ol’ days. That’s how the money boys see it anyway.
Do you know what good mass transit could do though? Imagine cities without parking lots and garages. Imagine having spaces that are much safer and more comfortable to walk. Imagine solving the housing crisis, since you can now build downtown complexes where those parking garages were.
Imagine getting most semis off the road and reducing road repairs by more than half.
Trains could do a lot, and it doesn’t take much imagination.
I mean there was a train bubble once, simpler times tho
We don’t need to imagine, we only need to look at China and see with our own eyes.
