What are we not getting in exchange?

The article The AI boom is causing shortages everywhere else (Hacker News) includes the following comment that attempts to not only put the capital expenditure in AI technology into context but also describes the immense opportunity cost.

“It’s hard to comprehend the scale of these investments. Comparing them to notable industrial projects, it’s almost unbelievable. Every week in 2026 Google will pay for the cost of a Burj Khalifa. Amazon for a Wembley Stadium.

“Facebook will spend a France-England tunnel every month.”

“ As a research topic, modern AI is a miracle, and I absolutely love learning about it. As an economic endeavor, it just feels insane. How many hospitals, roads, houses, machine shops, biomanufacturing facilities, parks, forests, laboratories, etc. could we build with the money we’re spending on pretraining models that we throw away next quarter?”

This is madness, of course, but it’s par for the course: some of us have observed the money machine pouring capital expenditure into military-hardware companies for a long time now, always to the detriment of any social investments, like those listed in the comment above: schools, daycares, public transport, “hospitals, roads, houses, machine shops, biomanufacturing facilities, parks, laboratories,” etc. That our elites are spending money on things beneficial to them rather than us is not new but that they’ve started wasting it on AI is.

Examining opportunity cost

Another commenter wrote that they’d,

“[…] just made a LLM recreate a decent approximation of the file system browser from the movie Hackers (similar to the SGI one from Jurassic park) in about 10 minutes. At work I’ve had it do useful features and bug fixes daily for a solid week.”

Implicit in this argument is that this person having recreated a file-system browser from a movie for pure fun—and with little to no effort on their part—did so in the context of a society that actually considered the costs, and billed the appropriate parties. That is, the commentator probably spent a few bucks on it but that’s only because society is massively subsidizing things that make the richest people richer and the richest people are currently obsessed with AI as the thing that will make them richer. People like the commentator are currently under the umbrella and benefit from the self-aggrandizing activity of the rich, at least tangentially.

For the rest of us, we really should be thinking about the opportunity cost, and we should very much be wondering why we’ve decided—mostly implicitly—that this person gets to have an infrastructure for playing around with their hobbies—or for being slightly more efficienthaving slightly more fun at work—while people in need of hospitals or medical care can go and hold a bake sale[1] or start a GoFundMe. This is fine, apparently, especially if we don’t really think about it.

LLM output is still mostly meh.

On top of it being an a moral affront that our societies prioritize military hardware and the self-aggrandizing dreams of tech elites and crypto bros above anything even remotely useful to everyone else, the output of these tools is still just not very good. They seem to have given up on making it better, and instead focused their efforts on media campaigns, advertising, press releases, and shills, all of which is meant to make us lower our standards instead.

The results may kind of suck, but you barely had to lift a finger to get them, and you can easily convince your boss—who also doesn’t care about the results—that you’re working while the little circle spins on your laptop, so it’s a win-win-win, right?

For example, here’s a picture illustrating the point of the previous section.

Military Expenditures vs. Social Spending

Do you like that graphic? Does it feel appropriate? Because, by way of illustration, I generated that image with Microsoft Copilot when I couldn’t find anything matching “military versus social spending scales” on DuckDuckGo that wasn’t a pie chart. I could have gone with a pie chart but I kind of wanted the scales. I’m spoiled too. I’m under the umbrella.

The result is kind of … off, though, isn’t it?

1. There is no physicality to it; it’s not grounded; it kind of floats.
2. The shadow doesn’t represent the two platters.
3. The rocket isn’t sitting on anything.
4. The wires holding the social-spending platter don’t go anywhere.

It’s fine, I guess? The results have gotten better but this is still where we’re at in 2026, after nearly $1T of capex and untold amounts of subsidies for energy infrastructure and data-centers. Given all that, that image is a wildly mediocre result.

It’s better than what I could have come up with in the minute that it took to generate it. But is it good? Absolutely not. It’s better than nothing. That is what we’re trading social services for: better-than-nothing versions of things we didn’t really need. So, elites (like myself) can control vast amounts of resources in the cloud to generate useless gewgaws while people go hungry and without healthcare.

This is the bargain. It always has been.[2]

Baubles for the rich always beat food for the poor

It’s different this time

Another comment on the Hacker News post writes,

“Is it the beginning of the star trek ship computer? If so, it is as big as the smartphone, the internet, or even the invention of the microchip. And then the investments make sense in a way.”

This is the one. Can’t miss.

Keep telling yourself that, buddy.

People keep claiming that these tools will eventually turn around and solve all of the other societal problems, which is why it’s absolutely sensible, patriotic, and moral to put all of our eggs in exactly this basket, this time.

Forget about all the other times. Lord knows we’ve tried hard enough to get you to forget about all of the other times we’ve scammed you.

Even if you vaguely remember that something unpleasant might have happened once or twice—or might even vaguely remember who was responsible—just forget all that. Because this time, it’s different. This time it will work. Promise. Cross our shriveled little hearts and hope to die.

Seriously, don’t even worry your pretty little head about it: There is no way this will turn out to enrich all of the usual suspects, leaving the rest of us with nothing. No way. This is the one. This time it’s real.

The same assholes who already own everything are recruiting you into their propaganda campaign to increase their fortunes. They’re saying, once again, “Let’s just do this thing first, then we’ll get to all of the things you need. Don’t worry, we won’t forget you.”

Lucy is soooo convincing, isn't she?

Maybe LLMs will kill LARPing I dunno

When I read about people building five projects a week, or submitting 27 PRs a day, I’m reminded of people who say that they read 200 books a year. This may be superficially true, but they are almost certainly crap books, or they’re just skimming them, or they’re incapable of understanding them. They are cheating. They are rounding up. They are emphasizing quantity over quality, which, like, used to be a bad thing.

Now, the barrier has been lowered even farther. People can now write 50-page “papers.” They can write “full-fledged apps.” Because the barrier to entry has been drastically lowered, there is less room for those LARPing as writers or programmers these days, not because they can’t LARP anymore but because no-one can tell the difference between their LARPing and LLM-supported LARPing.

A dozen years ago, the doors were wide open for people who could barely spell JavaScript—and had no idea what the difference was between that and Java—to earn six-figure salaries while building careers in an industry they had no hope of understanding. There was a lot of money sloshing around in the industry and managers greedily took up the slack in order to fill their teams with heartbeats.

These managers weren’t interested in actually accomplishing anything, but did it in order to look like they might accomplish something for long enough for them to get promoted like a space shuttle achieving orbit. They dropped their team like booster rockets, which careened back to Earth, only to be picked up by another enterprising manager more interested in a career than in actually accomplishing anything.

This worked out great for everyone as long as the industry was awash in money for such escapades. It no longer is, as those with all of the money have moved on to playing much larger games that don’t involve minor cogs earning six-figure salaries. They are instead focused on landing ten-figure deals that also have no hope of ever providing any value outside of making them money, but that’s the play these days apparently.

Long story shot, the LARPers are now having a tough time of it. They LARPed for so long that they think that they’re actually engineers whose jobs are endangered by the new fad on the block, LLMs. No-one cares about SPAs anymore because people who don’t care whether products actually work over the medium- or long-term can have any monkey churn out dozens of them a day to show their bosses. who also don’t care whether the products work because our entire economy is built on LARPing with only awards and no accountability or consequences in sight for the greatest transgressors.

Does it matter if anything even works anymore?

I think it’s premature to predict the end of anything when it’s completely unclear in what form any of what’s available today (A) will be available in that form and price point in the near future and (B) whether it even is what it claims to be—or what its most fervent acolytes claim it to be.

Hype is hype because it grows by repetition rather than by the introduction of new information. We are seeing a giant version of that and it feels inevitable because a lot of people are spending a lot of money to make it feel that way.

if you know what you’re doing, then you personally should have nothing to fear because you and I both know that the future will not be herding LLMs because it doesn’t work the way they say it works, nor will it until something significantly changes.

Since no-one seems to be interested in going anywhere near a drawing board to do some basic research, and since the amount of money being sloshed around to support the current LLM-based fantasy is larger than anything we’ve seen before, the aftermath is going to be epically bad.

I think that we can safely say that losing our jobs to AI will be the least of our concerns as we pick our way through the pillaged aisles of an abandoned grocery store in the post-apocalyptic hellscape that is definitely coming in the next financial crash that will make 2008 look like a bank error in their favor.[3]

Community chest bank error in your favor

Show me what you can do now

The verb case the proponents of this revolution use is always “future”. Success is just over the horizon. Just a little bit more. This is how MLMs work; it is not a serious business model.

These people love to round up. “We built a browser”. STFU. You did not. You built another prototype. You built another copy of something that already exists.

We are over four years into this mess and all I can see is software getting noticeably worse.

Hey, you jumped your bike over the ramp again, Billy. Cool. Can you do something useful yet? Like, can you go to the store and get me some goddamned cigarettes?

They have every incentive to lie bigly

All I see are fantasies spun by people worth hundreds of billions of dollars who run companies that are hundreds of billions of dollars in debt who are trying to keep the plates spinning so that you don’t notice that some of these people are planning to leave by the back door real soon.

When the CEO of Anthropic tells you that his company is going to change the entire world, it’s the same thing as when Trump says that polls no longer matter. They desperately need you to believe these things even though they don’t believe it themselves.

I think a great example of this is when Tesla quietly abandoned its autopilot program a little while ago, after years and years and years of telling people that they would be able to drive their own cars without touching the wheel—and after several people actually believed it so hard that they killed themselves in car accidents. Now, years later, that program is just completely gone. It is no longer officially a program just like it was never an actual non-imaginary thing to begin with.

That’s all a lot of this is. I’m happy if you find value in it for yourself. But it’s not here to provide value to you. It’s here to pretend to be much more valuable than it is so that a bunch of rich people can cash in, get out, and move on to the next scam.

I would have thought that a lot more people would have grokked that none of the people claiming that they’re here to save the world for all of us are even slightly interested in doing so, to say nothing of being capable of pushing us in that direction, if only a little bit.

Our radar for detecting grifters is completely broken, and that’s why society is saturated with them, top to bottom.