It means copilot will stay in your computer but do even worse, and cost your company 3 times as much.
askchapo
Ask Hexbear is the place to ask and answer ~~thought-provoking~~ questions.
Rules:
-
Posts must ask a question.
-
If the question asked is serious, answer seriously.
-
Questions where you want to learn more about socialism are allowed, but questions in bad faith are not.
-
Try !feedback@hexbear.net if you're having questions about regarding moderation, site policy, the site itself, development, volunteering or the mod team.
Right now your consumer usage is subsidised by venture capital funding, especially from Gulf States parking oil money. If they don't offer you more than the cost of token generation for free or near-free, you have plenty of competing models that do the same thing for free. If none of them are free, there are plenty of locally-run models that are. The AI bubble popping is when the investor funds dry up and the cost of token generation is passed onto consumers. It will try to follow the same enshittification process that streaming services have, but with a less valuable product. Trying to pass those costs onto businesses that are using AI to cut costs is so contradictory that the bottom falls out for the entire market.
Copilot won't leave your computer, it will just become like Office. I remember using Word and Excel for free but now they expect a subscription that's almost as obscene as Adobe's. Basic functionality of your computer will be paywalled to make up for the costs these companies sunk into AI bullshit, during the depression they caused where nobody can afford their products.
Copilot won't leave your computer, it will just become like Office. I remember using Word and Excel for free but now they expect a subscription that's almost as obscene as Adobe's
Right. Yep, I've already seen that at my current employer. Copilot is free for the time being but there is a strong push by Microsoft to get us onto an M365 paid Copilot licence at a nominal cost per user.
The only difference between the 'baked in' Office Copilot (which is literally now the Office 'apps' landing page) and the M365 copilot is that the latter purportedly will not use your tenant data to train models.
The app is functionally the same, and the only business units who want M365 paid version rolled out are governance and IT. All staff just use the free Copilot now or use other LLMs.
So if the app is functionally the same for user output, I guess the only way they can go is to remove the free one and have people clamour for the company to pay to get it back? So like, hopefully it may disappear from some parts of the Microsoft suite?
I've been using Windows since 3.1 and Microsoft stopped making good decisions after XP. They're such a bloated company that they're institutionally incapable of coming up with something new that's good, only doubling down on the things that have brand recognition. I could totally see them do something like make copilot run locally in the way Deepseek does but still charge a subscription fee to unlock half the functionality of Windows 12 with it.
I could totally see them do something like make copilot run locally in the way Deepseek does but still charge a subscription fee to unlock half the functionality of Windows 12 with it.
This sounds plausible. They'd still need to strike a balance so they can retain some of the data for training, presumably, so it wouldn't be 100% local? Who knows, they may do something similar with paywalled logic models that require online compute, and then hamstring the local only version. We'll see.
i mean 7 was nice.
7 was inoffensive and I'd go back to it over 11, but I liked Windows 7 because it felt like they undid all the Vista shit to restore Windows XP with a more modernised UI.
Copilot will never go away. The only way you can get that thing off your computer is by removing Windows. The AI bubble is gonna burst but Microsoft is still gonna have LLMs baked into their OS.
That makes sense. I'm not at liberty to change the OS on my work PC, I'm already on Linux on my home laptop. How about the absolute deluge of AI slop? Do you anticipate that as those companies/platforms go under (or there aren't the data centres to run the compute) that it will ameliorate in any way?
The AI slop is here to stay unfortunately. The bubble will pop and like 99% of AI companies will die, but LLMs are actually very easy to run locally AND at least one of the big AI companies will "win" and capture most of the AI traffic, kind of like how Google "won" search. As models get more efficient all those data centers are gonna be unnecessary, but they'll only become unnecessary because it's easy to run frontier models on your own phone or whatever. AI slop will never disappear, that's a Pandora's box that cannot be closed.
Understood.
As models get more efficient all those data centers are gonna be unnecessary, but they'll only become unnecessary because it's easy to run frontier models on your own phone or whatever.
I've read commentary here that the excess compute will go towards AI analysis for surveillance (to a greater degree than it already is), but I'm guessing there will be many un- and under-utilised data centres. Maybe we can turn them into ice skating rinks, indoor ski slopes and swimming centres after the revolution.
I'm guessing there will be many un- and under-utilised data centres.
I would argue that cloud computing is a viable model, and it would be comparatively easy to flip an AI data center for other uses.
I think the reason all the AI companies want intelligence & defense contracts is a) they can claim they're essential, and should get preferential bailouts, and b) surveillance/intelligence is basically an infinite market for the excess capacity they'll be left with after the crash.
It's going to become a new Bonzi Buddy, but it's going to be embedded into the kernel.
The corn kernel, yim yum 
It means Copilot will be abandoned by Microsoft but will still be on every PC and be used as a backdoor for easy hacking.
Much of this is just speculation, but still…
What you should understand is that the technology isn’t going to go away. Chatbots generally won’t be going away. What we’ll likely see is consolidation and bailouts. In a perfect world, the companies would just go under, but governments have shown a total willingness to buy into the technology.
All a bubble really is, is a market or industry that is overvalued due to speculation. What we will likely see is the biggest companies swallowing up the smaller ones and AI technology becoming some kind of software-as-a-service offered by a handful of companies. Microsoft, Google, OpenAI… We can’t know which companies, and that’s part of what is keeping this bubble going. I know I’m oversimplifying, but we’re not talking about what caused the bubble, just what will happen when it pops. The shoehorning won’t stop, but it will become less obnoxious. The tech will find its place to live, and in some of those places, it won’t seem that bad. I’d guess in the form of an AI agent assistant.
I think Microsoft’s dream model is actually the future that we’re looking at. That is, some kind of AI-integrated OS or web browser (or both) that is feeding you information as filtered through the AI. Meaning all information you get from the internet would be fed to you, not directly from the host source, but from whatever company runs your AI OS or browser. That’s a little theoretical right now, granted, but the dream seems to be that for these companies.
AI integration is garbage, but at the front end, your average user isn’t going to know that. I suspect that’s how they wanted it. We are seeing the foundations of a redesign of the internet and the digital age in real time. It’s not just companies wanting to get a return on value from these LLMs. It’s also the gamble that when the various competitors collapse, they’ll be able to swoop in, grab the competing tech for pennies, and implement it themselves.
When the bubble pops, we will likely see huge collapses in the tech sector as all the companies that invested in it directly get swallowed up. Something something too big to fail. Some of those sectors will get bailed out, and the US will likely suffer a massive depression from the various costs. There is concern about the potential for AI being used to create a software and hardware-as-a-service model, and I can see that being something that gets pushed. Because then, the data center infrastructure that is getting built serves a new purpose.
So it’s GOING to pop, but "deflate" might be more accurate. I know that most of us don’t want it and that it’s pretty hard to see right now, but all the pieces of the puzzle are present or being built.
I don’t actually see much changing on the hardware side. Computer components manufacturers have all been shifting to producing for AI data centers. So anyway, let me make bullet points of my thoughts because all this is messy.
- AI doesn’t go away; it becomes the new front end of the internet and home computing.
- AI companies consolidate and maybe specialize into various parts of the market (think Google offering web browsing, and Windows offering home computing).
- Home computing will become virtualized, with the home computer becoming something of an access point for a rented VM, which meets their needs based on what they pay for. We’ve seen this fail with cloud gaming, but only because there are still other options for access to things like games and software.
- Access to information is controlled and obfuscated as websites and hosting become more expensive. (This is already happening now, but with the consumer market shrinking for computer parts, the only viable option may be for users to buy into this model.)
There’s more, but really, that’s my big picture, and why pretty much everyone ignores the general public's hate for it. They don't see it as a fun new fade, they see it as an innovation that his going to take over the world. as much as I hate it I think they're right, If only because at the same time that all of this is happening we are seeing industry captures through legislation and corruption.
The bubble popping is going to hurt consumer markets not commerical ones and all the tech they've developed and hardware they've developed will be used to help absorb costs and take over those markets which the bubble itself destroyed. We'll all suffer for it.


Thank you for the thorough writeup, it sounds like it's a very probable outcome. Deflate is a good word for it. I think last I stepped off reading about the AI bubble it was an Ed Zitron, Zircon? (or whatever his name is) which was quite bombastic about the whole AI house of cards falling, but light on about the 'well, what about after the bubble pops' doubts/questions I had lingering in my mind
Yeah this fleshes out my speculation as well. It was more about rearranging the infrastructure. Uh, I better start getting really good at accessing library records before they wall that stuff off too
I find that Gamers Nexus on YouTube does really good reporting on it from the hardware side of things. Most of their reports are about an hour long and take quotes directly from primary sources. It’s hard to give other primary sources beyond that. I tend to get financial news from all over, and I enjoy learning about history.
It won't be as clean and permanent as the NFT crash was. Hard to say for sure if you'll get the specific wins you're looking for.
It's a lose-lose proposition whether the bottom of it falls out or not. The investment in AI is the only healthy looking "growth" sector in our economy. So if you don't want what feels like the looming recession to happen, then you want this fantasy to continue. But this fantasy is built on this evil technology.
And so when the bubble pops, and suddenly everyone has to pay for this shit by way of enduring a recession, at least that will make using AI itself less palatable because it won't be available for its current highly subsidized pricing, but those unemployment rates and prices on things will likely be going up.
But the technology isn't going away, much like how the atomic bomb isn't going away. It exists. It's part of reality now.
LLMs have some uses, NFTs were pure fiction. AI will definitely deflate, but I don't see the tech going anywhere. It'll probably become way more expensive tho.
Blockchain has a use and all the companies who were trying to jam blockchain into irrelevant shit are gone. I think AI companies shifting away from grabbing eyeballs to actually making money will help raise prices of enterprise licenses so more companies stop forcing it on all their employees and start rationing it. Right now our fucking chats are fed into it so when everyone signs off for the night, we get a notification that says something like "various participants wished the others a pleasant evening." That might cost a penny now but it made somebody feel like they were on the cutting edge of something. When it costs a dollar and is clearly going to remain precisely that useless, I sure hope that same somebody pulls the plug.
Blockchains aren't necessarily good or efficient things, but they do still exist. Unlike NFTs. I don't have any reason to think AI won't be the same. Transformers and LLMs are way better at NLP than anything else that has been invented so far. That's not going away, people will just have to actually justify their use once the prices start to match the costs.
That said, I think looped models, RYS, and other alt model architectures will continue to drive the training costs down like Deepseek did and totally demolish the AI company valuations. Just make it bigger was never going to be a winning strategy, it's completely naive. That's really the source of the bubble. The technology has nothing to do with it, the tech bros are just desperate for a machine god and they got caught up in their chat bot delulu about it and made it everyone's problem.