> There are many technologies that have seemed inevitable and seen retreats under the lack of commensurate business return (the supersonic jetliner)
I think this is a great analogy, not just to the current state of AI, but maybe even computers and the internet in general.
Supersonic transports must've seemed amazing, inevitable, and maybe even obvious to anyone alive at the time of their debut. But hiding under that amazing tech was a whole host of problems that were just not solvable with the technology of the era, let alone a profitable business model. I wonder if computers and the internet are following a similar trajectory to aerospace. Maybe we've basically peaked, and all that's left are optimizations around cost, efficiency, distribution, or convenience.
If you time traveled back to the 1970s and talked to most adults, they would have witnessed aerospace go from loud, smelly, and dangerous prop planes to the 707, 747 and Concorde. They would've witnessed the moon landings and were seeing the development of the Space Shuttle. I bet they would call you crazy if you told this person that 50 years later, in 2025, there would be no more supersonic commercial airliners, commercial aviation would basically look the same except more annoying, and also that we haven't been back to the moon. In the previous 50 years we went from the Wright Brothers to the 707! So maybe in 2075 we'll all be watching documentaries about LLMs (maybe even on our phones or laptops that look basically the same), and reminiscing about the mid-2020s and wondering why what seemed to be such a promising technology disappeared almost entirely.
I think this is both right and wrong. There was a good book that came out probably 15 years ago about how technology never stops in aggregate, but individual technologies tend to grow quickly and then stall. Airplane jets were one example in the book. The reason why I partially note this as wrong is that even in the 70s people recognized that supersonic travel had real concrete issues with no solution in sight. I don't think LLMs share that characteristic today.
A better example, also in the book, are skyscrapers. Each year they grew and new ones were taller than the ones last year. The ability to build them and traverse them increased each year with new technologies to support it. There wasn't a general consensus around issues that would stop growth (except at more extremes like air pressure). But the growth did stop. No one even has expectations of taller skyscrapers any more.
LLMs may fail to advance, but not because of any consensus reason that exists today. And it maybe that they serve their purpose to build something on top of them which ends up being far more revolutionary than LLMs. This is more like the path of electricity -- electricity in itself isn't that exciting nowadays, but almost every piece of technology built uses it.
I fundamentally find it odd that people seem so against AI. I get the potential dystopian future, which I also don't want. But the more mundane annoyance seems odd to me.
> even in the 70s people recognized that supersonic travel had real concrete issues with no solution in sight. I don't think LLMs share that characteristic today
I think they pretty strongly do
The solution seems to be "just lower your standards for acceptable margin of error to whatever the LLM is capable of producing" which should be concerning and absolutely unacceptable to anyone calling themselves an Engineer
99% or more of software developers behave in ways that would be inconceivable in actual engineering. That's not to say there aren't software engineers, but most developers aren't engineers and aren't held to that standard.
“Increasing Success by Lowering Expectations” That is from Despair Inc.
This was obviously meant to be funny by them, now it looks like the state of play.
> The reason why I partially note this as wrong is that even in the 70s people recognized that supersonic travel had real concrete issues with no solution in sight. I don't think LLMs share that characteristic today.
The fundamental problem has already been mentioned: Nobody can figure out how to SELL it. Because few people are buying it.
It's useful for aggregation and summarization of large amounts of text, but it's not trustworthy. A good summary decreases noise and amplifies signal. LLMs don't do that. Without the capability to validate the output, it's not really generating output of lasting value. It's just a slightly better search engine.
It feels like, fundamentally, the primary invention here is teaching computers that it's okay to be wrong as long as you're convincing. That's very useful for propaganda or less savory aspects of business, but it's less useful for actual communication.
I think what you meant to say is that costs are high so they can't generate large profits. but saying that they can't figure out how to sell it seems absurd. Is it Netflix level of subscribers, no. But there can't be more than a couple of hundred products that have that type of subscription reach.
Ok but isn’t 20 million subscribers out of what, 800 million or 1 billion monthly users or whatever they’re claiming, an absolutely abysmal conversion rate? Especially given that the industry and media have been proclaiming this as somewhere between the internet and the industrial revolution in terms of impact and advancement? Why can they not get more than 3% of users to convert to paying subscribers for such a supposedly world changing technology, even with a massive subsidy?
As another commenter notes, because you get access to a lot of functionality for free. And other providers are also providing free alternatives. The ratio for their free/paid tier is about the same as YouTube's. And like YouTube, it's not that YouTube isn't providing great value, but rather that most people get what they need out of the free tier.
The better question is what if all LLM services stopped providing for free at all -- how many paid users would there then be?
A service like Gmail or Dropbox with low storage is close to free to operate. Same thing with iCloud - 50 gigs a month is what, 1 dollar? How is that possible?
Because 50 gigs is next to nothing, and you only need a rinky dink amount of compute to write files.
YouTube, on the other hand, is actually pretty expensive to operate. Takes a lot of storage to store videos, never mind handling uploads. But then streaming video? Man, the amount of bandwidth required for that makes file syncing look like nothing. I mean, how often does a single customer watch a YouTube video? And then, how often do people download files from Dropbox? It's orders of magnitude in difference.
But LLMs outshine both. They require stupid amounts of compute to run.
Close to free per user, maybe. But dropbox has 800 million users, only ~2% pay, and Gmail has billions. They spend a lot of money running those services.
They are purposely losing billions, this is a growth phase where all of the big AI companies are racing to grow their userbase, later down the line they will monetize that captured userbase.
This is very similar to Uber which lost money for 14 years before becoming profitable, but with significantly more upside.
Investors see the growth, user stickiness and potential for the tech; and are throwing money to burn to be part of the winning team, which will turn on the money switch on that userbase down the line.
The biggest companies and investors in the planet aren't all bad at business.
I'd say the userbase has grown. You can't claim half a billion users and simultaneously say you're still trying to grow. This isn't a month-old technology now. And they still can't turn a profit. (edit: and by "you" i meant "they")
>You can't claim half a billion users and simultaneously say you're still trying to grow.
You can if you're still growing. ChatGPT is the 5th most visited site on the planet yes, but it is still growing hundreds of millions of visits with every passing month.
They aren't turning a profit because they aren't monetizing the vast majority of subscribers in any way (not even ads). LLM inference is cheap enough for ads to be viable.
In my companies, AI subscriptions and API access are now the biggest costs after salaries and taxes. Don't know what makes you think these services aren't attracting paid customers?
> even in the 70s people recognized that supersonic travel had real concrete issues with no solution in sight. I don't think LLMs share that characteristic today
I hate to dogpile on this statement but I can think of two major issues right now:
* Small context windows, and serious degradation when pushing the limits of existing context windows. A human can add large amounts of state to their "context window" every day.
* Realtime learning. My humans get smarter every day, especially in the context of working with a specific codebase.
Maybe the AI companies will figure this out, but they are not "same technique more processor power" kinds of problems.
There are sound math reasons for skyscrapers topping out, mostly due to elevator capacity and the inability to effectively get people in and out of the floorspace as you go past a few hundred ft. There's no construction engineering reason you can't go taller - the Burj Khalifa, for example, is three times taller than a typical Western major city skyscraper - it just doesn't make economic sense unless you're a newly rich nation looking to prove a point.
Economic Concrete construction (what China specializes in) typically tops out at 30-40 floors, so the vast majority of buildings in Asia are that height, a sweet spot so to speak especially for residential (even in limited space HK).
>I think this is both right and wrong. There was a good book that came out probably 15 years ago about how technology never stops in aggregate, but individual technologies tend to grow quickly and then stall. Airplane jets were one example in the book. The reason why I partially note this as wrong is that even in the 70s people recognized that supersonic travel had real concrete issues with no solution in sight. I don't think LLMs share that characteristic today.
I don't see any solution to hallucinations, nor do I see any solution in sight. I think that could count as a concrete issue that would stop them.
Vision and everyday-physics models are the answer: hallucinations will stop when the models stop thinking in words and start thinking in physical reality.
They had easy access to a large corpus of writing to train on, way larger than any human being trained their own language model on. I can't see where they are going to find a large corpus of physical interaction with reality to train that kind of model.
Yeah, and with LLMs the thing I can't shake, however, is that this time it's pretty strongly (maybe parasitically) latched onto the aggregate progress of Moore's law. Few other technologies have enjoyed such relatively unfettered exponential improvement. It's like if skyscraper materials double in strength every n years, and their elevators approach teleportation speed, the water pumps get twice as powerful, etc., which would change the economics vs the reality that most of the physical world doesn't improve that fast.
Was the problem that supersonic flight was expensive and the amount of customers willing to pay the price was even lower than the number of customers that could even if they wanted to?
- They were loud (sonic booms were nasty).
- They were expensive to maintain and operate. Guzzlers. (Britain and France clung to them as a matter of pride/ego)
- They were narrow and uncomfortable. I have seen videos where there is space only for one stewardess to walk. I had been inside of one in Seattle museum. Very cramped.
- As you mentioned, ticket cost was high.
- I suspect people traveled in these mostly for bragging rights.
You made this point in passing, but it's so relevant to LLMs I wanted to highlight it: The development and operational cost was heavily subsidized by the British and French governments, because having an SST was a point of national prestige.
Yeah, basically. Nobody wanted to pay $12,000 to be in a plane for three hours when they could pay ~$1200 to be in one for six hours. Plus, they used up a lot of fuel. That made them real vulnerable to oil price spikes.
Contrast that with modern widebody jets, which fly ~300 people plus paid cargo on much more fuel-efficient engines.
From a system optimisation perspective, SSTs solved the wrong problem.
Want to save people time flying? Solve the grotesque inefficiency pit that is airport transit and check-in.
Like, I'm sorry, STILL no high speed, direct to terminal rail at JFK, LAX and a dozen other major international airports? And that's before we get to the absolute joke of "border security" and luggage check-in.
Sure, supersonic afterburning engines are dope. But it's like some 10GHz single-core CPU that pulls 1.2kW out of the wall. Like it or not, an iPhone 16 delivers far more compute utility in far more scenarios.
It makes it all the dumber that we have the tech and still can't manage to implement the solution.
Like an org with crappy management and team structure shipping bloated, buggy code even though they've the budget to hire great engineers and the problems they're solving are largely known and well-trodden.
It did for international, maybe not at the dawn of SSTs but after a string of hijackings in the 70s/80s they brought it in. Not for US internal flights, it's true.
The crucial point is that we simply do not know yet if there is an inherent limitation in the reasoning capabilities of LLMs, and if so whether we are currently near to pushing up against them. It seems clear that American firms are still going to increase the amount of compute by a lot more (with projects like the Stargate factory), so time will tell if that is the only bottleneck to further progress. There might also still be methodological innovations that can push capabilities further.
> So maybe in 2075 we'll all be watching documentaries about LLMs (maybe even on our phones or laptops that look basically the same), and reminiscing about the mid-2020s and wondering why what seemed to be such a promising technology disappeared almost entirely.
It's hard for me to believe that anyone who works with technology in general, and LLMs in particular, could think this.
slower, no fast option, no smoking in the cabins, less leg room, but with TVs plastered on the back of every chair, sometimes
its actually kind of scary to think of a world where generative AI in the cloud goes away due to costs, in favor of some other lesser chimera version that can't currently be predicted
but good news is that locally run generative AI is still getting better and better with fewer and fewer resources consumed to use
The problem with supersonic commercial jets was mainly one of marketing/politics. The so called "sonic boom" problem was vastly overhyped, as anyone who lives near an air force base can tell you.
The conspiracy theorist tells me the American aerospace manufacturers at the time (Boening, McDonnell-Douglas, etc.), did everything they could to kill the Concorde. With limited flyable routes (NYC and DC to Paris and London I think were the only ones), the financials didn't make sense. If overland routes were available, especially opening up LA, San Francisco and Chicago, it might have been a different story.
>as anyone who lives near an air force base can tell you.
In the US, the Air Force is simply not allowed to fly supersonic anywhere near a city or a suburb with only a few exceptions.
One exception is Edwards Air Force Base in the California desert: there are houses nearby, but the base (and supersonic warplanes) preceded the construction of the homes, so the reasoning is that the home builders and home buyers knew what they were buying into.
Another exception (quoting Google Gemini):
>From 1964 to 1966, the FAA and U.S. Air Force conducted supersonic flights over
St. Louis and other cities like Oklahoma City to gauge public reaction to daily
sonic booms. The goal was to understand public tolerance for commercial
supersonic transport (SST) operations. Reactions in St. Louis, as elsewhere,
were largely negative, contributing to the eventual ban on commercial
supersonic flight over land in the U.S.
Have you have experienced sonic booms? I have (when my family visited West Germany in 1970) and I certainly would not want to be subjected to them regularly.
Seems... wrong. Booms broke windows and drove zillions of complaints. Supersonic flight near airbases is controlled and happens on specific traffic corridors, right?
I think this is a great analogy, not just to the current state of AI, but maybe even computers and the internet in general.
Supersonic transports must've seemed amazing, inevitable, and maybe even obvious to anyone alive at the time of their debut. But hiding under that amazing tech was a whole host of problems that were just not solvable with the technology of the era, let alone a profitable business model. I wonder if computers and the internet are following a similar trajectory to aerospace. Maybe we've basically peaked, and all that's left are optimizations around cost, efficiency, distribution, or convenience.
If you time traveled back to the 1970s and talked to most adults, they would have witnessed aerospace go from loud, smelly, and dangerous prop planes to the 707, 747 and Concorde. They would've witnessed the moon landings and were seeing the development of the Space Shuttle. I bet they would call you crazy if you told this person that 50 years later, in 2025, there would be no more supersonic commercial airliners, commercial aviation would basically look the same except more annoying, and also that we haven't been back to the moon. In the previous 50 years we went from the Wright Brothers to the 707! So maybe in 2075 we'll all be watching documentaries about LLMs (maybe even on our phones or laptops that look basically the same), and reminiscing about the mid-2020s and wondering why what seemed to be such a promising technology disappeared almost entirely.