this post was submitted on 09 Oct 2026
936 points (97.7% liked)
Technology
88725 readers
2886 users here now
This is a most excellent place for technology news and articles.
Our Rules
- Follow the lemmy.world rules.
- Only tech related news or articles.
- Be excellent to each other!
- Mod approved content bots can post up to 10 articles per day.
- Threads asking for personal tech support may be deleted.
- Politics threads may be removed.
- No memes allowed as posts, OK to post as comments.
- Only approved bots from the list below, this includes using AI responses and summaries. To ask if your bot can be added please contact a mod.
- Check for duplicates before posting, duplicates may be removed
- Accounts 7 days and younger will have their posts automatically removed.
Approved Bots
founded 3 years ago
MODERATORS
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
No. Demand for OpenAI's services is lower than they thought.
Every day, the open weights models get smarter and cheaper and new competitors are coming online every few weeks. They cost a fraction of what OpenAI charges.
Demand for AI products and services is continuing to grow at a break-the-world pace. It's just these big "AI first movers" are in trouble because it's too easy to copy what they do. They expected an enormous "first mover advantage" (which is a real thing) but it turns out that AI is trivial to train and replicate (just have to scrape the Internet and scan a lot of books).
In other words, it's turning out that "first mover" is a disadvantage in this (AI) industry.
I wouldn’t believe that hype
I am the only person in my circle using open weight models.
Normies out number us by a large amount and none of them really are running their own things.
Individuals using AI are a tiny drop in the bucket. Think bigger: Businesses and government.
There's government agencies racking up token usages to the point where Google had to build an entire data center on Christmas Island (the place with the cool crabs) and that still won't be enough! And that's just one government!
Companies are using AI in automations now and it's working so well that they're willing to spend tens of thousands of dollars a month (on tokens) just on that one automation.
These usages are where demand is growing. Not regular people or even developers using AI to help them write code.
Citation needed. I work in a midsized IT corpo, I don't see it working that well? Sure, some shit is shipped faster, some is shipped slower, customers churn is 2x higher than "planned" and prod bugs have doubled.
(And we do all the fancy shit, looping, custom harnesses, orchestrations, handovers, savepoints, post delivery pipelines, autoremediation, blablabla, context, graphs blabla)
Yeah that's not what I'm talking about. That's just coding stuff and I think everyone is learning a hard lesson right now that there's really two ways to do AI (for code):
#1 is a very expensive way to do it but it seems to work, albeit with a huge learning curve and a much longer horizon before you get something stable. Even then, what you get can be a giant fucking mess that's really, really hard to understand.
#2 is much safer and works quite well, IMHO. When the AI fucks something up, it's only a tiny little thing that's easy to fix. This method really is just a way for developers to improve their productivity.
With #2, it only takes a small amount of tokens to have the AI help you troubleshoot as well. At least, that's been my experience.
Having said that, the places where we are seeing the most growth in AI is small shit. For example, at my work they implemented an agent that scans customer complaints and filters out false positives. I work for a huge company that has a zillion products/services and the AI has tool calls into our ticketing systems to check if there was an outage at the time the customer complained. It's not that smart, but it's smart enough to tell if a ticket about system X was opened when system X had an outage. It can even tell if the customer was complaining about something relevant to the outage.
Stuff like that runs 24/7 and eats up bazillions of tokens, but only on our local hardware (we have thousands of enterprise GPUs). At other companies, they're doing similar things and slowly realizing they can replace OpenAI/Anthropic API calls with local ollama stuff.
Even if your token bill is $10,000/month for a service like that, it's still cheaper than paying an entire team of humans to perform the same function.
That is the growing use of AI automation I was talking about.
I feel like in a few years we will look back on this whole era and realise that our politicians and business leaders fell for a not so elaborate scam. LLM usage will be severely curtailed because of the scammy stank on them and they will slowly migrate back into the niches they should have been confined to in the first place.
Rightly or wrongly people got scared off widespread nuclear after Chernobyl and Three Mile Island, LLMs will be much the same. The damage they have done to economies will take decades to resolve and it will be interesting to see if the US public can muster the unity and wherewithal to hold the oligarchs, kleptocrats and their enablers responsible.
Even more interesting if some laws can get put in place to limit the powers of the executive and or introduce some safeguards with teeth to prevent this sort of catastrophe again.
Imagine if worldwide we treated the criminals with power and money like criminals.
I fear we will have a generational loss of intellectual from people who have rapidly learned to farm all their higher cognition off to the liar box, and then got excited and cheered about the liar box getting better at lying. But hopefully we can move past this.
Of course the other scenario is that China waits until the US economy is so intrinsically linked to LLMs that there is no extrication, then they start pulling investment. I guarantee you the Chinese government will remain in control of their population longer than the US government controls theirs once an economic freefall kicks off towards a recession that makes the great depression look like a blip in the accounts books.
It's too late for that scenario. LLMs are already widespread on millions of computers.
On my old 4060 Ti 16GB I can run Qwen3.8:14b and that's actually good enough for loads and loads of things. It's actually overkill for a lot of stuff too which is why tiny models are all the rage right now (e.g. 0.8b models).
Let's say the government banned LLMs. Are they really going to want to halt pharmaceutical and biomedical research like that? Are they going to start inspecting/requiring everyone's computers be scanned regularly for AI models?
No. It's unrealistic to think they can do anything at all to stop AI. If we ban it in the US, the world will just use AI services offered from other countries and then the US won't be able to compete because those countries will become vastly more productive.
As much as everyone on Lemmy believes all AI output is "slop", it really does do useful stuff! It's just most of the useful stuff doesn't make the news. It's boring things like checking if an uploaded image is NSFW or if an uploaded document has everything it's supposed to have or follows the correct format.
AI/LLMs are also checking the grammar of basically everything that gets published. Anyone who's not using AI for that is an idiot. It's just way too fucking good at spotting grammatical errors, incorrect word use (often innocent typos that normal grammar checking tools miss), and similar problems.
There's fucktons of super useful, good ways to use AI. We'd be stupid to just ban it and pretend it isn't good for a lot of things.
We can regulate its use but that's not really "regulating AI" as much as it is "regulating capitalism" or "regulating people". Because the AI isn't just going out on its own to spam misinformation or mislead consumers with generated product photos. Someone told it to do those things.
What we need isn't "AI regulation" is regular regulation that prevents companies from behaving badly.
Predictable as clockwork... Sure, there's no way we say, "fuck this shit is cancer maybe I should stop handing my cognitive capabilities over to a machine and start thinking again."
I had a whole big screed written out about the myriad ways that the implementation of this technology is corrosive at a societal level. But it's all just so exhausting to have the same arguments time and again.
So let's just leave it at, sure buddy, you and your ilk have the mandate of heaven to topple the world. Have fun, but please don't be surprised where the road you're on takes you.
No, no. You've got the wrong idea.
Will AI completely ruin a zillion people's ability to think (beyond what social media already accomplished)? Quite possibly!
I'm just saying there's nothing we can do to stop it at this point.
You're about to head down a long, lonely road as normies start using and accepting AI more and more into their lives while you scream into the abyss. I've been there with fossil fuels, copyright, software patents, and Linux.
People saying "Just use AI!" Is about to live rent free in your head as your most-hated phrase. Just like, "stop driving gas guzzlers!" Or, "stop bitching about Windows and just install Linux already!" Lives in so many others.
At least most people don't think of AI as a core part of their identity (e.g. pickup trucks) 🤷
Eventually we are bound to take steps to mitigate it. Like Radium, or Asbestos, or Lead in bloody everything, or cigarettes. Eventually the winds of change blow and the shitheels in power pivot to something new. It will be imperfect and it won't reach everywhere and we will be pulling it out of places we never suspected we would find it in 50 years.
I can even imagine some far fetched scenarios where people start seeing Dune as a prophetic vision, especially with the idea of LLMs targeting weaponry in conflict zones. We seems to have a fairly accepting view on people committing genocide... Will we be so open minded about LLMs doing a bit of genocide... Especially if it's gets its target wrong, maybe engages in a bit of friendly genocide on the aggressor countries own civilian populace.
I don't think it'll happen like Radium or Asbestos because those rely on scarcity. They're physical things that can be physically controlled.
AI models aren't like that. In less than a minute anyone can download the latest models. In under an hour, they can have all the latest models sitting on their PC. Right now, there's a shortage of memory and GPUs which makes them prohibitively expensive for the average person but that won't always be the case!
Ten years from now, everyone's phones will have enough capabilities to run models that currently require a PC with a 32GB GPU.
What we could do is force AI companies and software to put warnings in the chat interfaces like cigarette cartons: "WARNING: Using LLMs for just ten minutes increases the likelihood that you will have trouble completing difficult mental tasks afterwards." (That was in the news this morning)
We also definitely need international treaties that prevent countries from allowing AI to make the decision to kill people or destroy infrastructure. Though honestly, we're already in a state where countries are willing to kill hundreds of people just to assassinate a single target (e.g. US and Israel bombing an entire building because they think their target is inside).
These are small potatoes, though, compared to what AI could do to capitalism and inequality. The rate at which it is advancing is faster than people/jobs can keep up. It's like half the population is ferriers and suddenly everyone was just given a free car.
Ahh to be a techno optimist.
Things could get a whole lot more authoritarian than you think. It's very early here and I can't be bothered writing out the whole response, but we have governments around the world currently trying to make using a computer require ID. We are thin edge of the wedge here with age verification to "protect the kids". Looked at from an authoritarian perspective a country with the power to dictate terms of trade says LLMs are actually military tech and need to be controlled, implement laws to restrict them or else...
If your country having access to international markets means you lose access to certain software... Now let's complicate the situation by assuming a bubble burst leads to the world heading into a mass depression, do you really think the US has the wherewithal to retain its top spot? Times of mass political and financial upheaval can, and arguably almost always do, lead to radical changes to the living circumstances of people.
In 10 years time your phone won't be running shit that the government/corporation in control doesn't explicitly approve of and monitor.
It's not fucking AI. Stop repeating the marketing garbage, and stop advocating for slop.
What isn't AI? What are you even talking about?
BTW: I don't work for any AI companies and I don't believe the hype for one second that they're building extinction-level AGI. What they're building is general-purpose "intelligence" that's OK-ish at some things, great at others (that don't make them much money, LOL), and fucking terrible at anything that requires reliable truth (on the first try, LOL).