Just drop the price by 90% in a year and a half while already hemorrhaging money, no biggie. haha.
If these stupid fucks didn't have the ability to affect my life this would be so, so funny. It's still pretty funny even so.
Just drop the price by 90% in a year and a half while already hemorrhaging money, no biggie. haha.
If these stupid fucks didn't have the ability to affect my life this would be so, so funny. It's still pretty funny even so.
This is so deeply, fundamentally hilarious to me.
I am genuinely convinced we’re gonna see the bubble pop before the end of the year.
I really, really hope that within 1 or 2 years there will be a chain of collapses of companies which adopted this shit.
People in those companies economies (and others globally to a lesser extent) will feel the economic fallout from that in one form or another
Those companies fired the workers and replaced them with AI already, so the benefit they bring to the broader economy is far smaller (might even be negative by now) and them dissapearing is far less likely to cause "economic fallout".
Even though that's what Neoliberal politicians have preached to us for decades now, companies aren't inherently "good for the Economy" - they might, but then again they might not.
Plenty of companies out there add up to a negative to the Economy:
Companies aren't there to benefit the Economy in any way form or shape, they're there to benefit their owners. Some companies might happen to benefit the Economy, often because they have no other option (for example, they're forced to employ people in order to actually operate and make money for the company owners, thus indirectly help the Economy) but almost invariably every single thing that helps the Economy is for a company "losing money" which would go to the company owners instead so they generally fight like crazy to actually reduce it: they try to spend less from suppliers, pay less to workers, pay less or no tax at all.
Helping the Economy almost always runs counter the one true goal for the existence of a company - maximizing shareholder profits.
Further, when a company dissapears it doesn't mean that the Market need that they're serving dissapears with them: in almost all the cases the need is still there after they've collapsed (unless they collapsed because the need itself doesn't exist anymore) so existing companies grow or new companies pop-up to serve it.
It seems to me that companies that have replaced workers with AI are by far the most likely to by now not actually be helping the Economy, not even indirectly, and that their dissapearance might actually improve the Economy if only because their share of the Market might be taken over by companies with better far better and less wasteful management (and hence more efficient) and which employ more people, and hence which bring more benefits to the broader Economy.
I don't disagree with any of you're points. I guess I was more getting at the idea of the seemingly more ever present AI bubble collapse (which I am very much in favour of) will come with some pretty serious down sides for ordinary people (which I'd like to clarify is the fault of the AI industry at large).
I guess I get a bit puzzled when people cheer on the popping of said bubble because I can't see good things happening.
Economically, the "AI" to where most investment money went to - LLMs - has turned out to be one big value destroying ponzi scheme, hence removing it from the Economy will in overall improve the Economy because things are actually more efficiently done without it than with it.
The area were LLMs shine is entertainement, and in overall Economic terms entertainment can be a good thing or a bad thing - on one side it's a use of time that is not economically productive when said time could instead have been used for productive activities, on the other hand it can indirectly boost other Economic activities because humans work better when they're more relaxed and satisfied. From a lot of what we're seeing, LLMs not only are misused to do work whilst in fact being entertainment (for example, CEOs using LLMs for "research" because it's the ultimate "Bootlicker Yes man" even though its highly unreliable as a research tool, thus leading to a higher rate of bad management decisions) but it also seems to reduce independent thinking capabilities of its regular users.
The second biggest sink of investment money in what is now called "AI" is Generative AI. This comes with job losses and a fall in quality (but it's in areas which don't directly create wealth, hence the Economic impact of that loss in quality is lower). Ultimatelly, the reversion of this area of "AI" being "good for the Economy" or "bad for the Economy" boils down to what one thinks is a "good Economy" - if one sees inherent Economic value in Art for itself, then a fall in the quality of Art and a reduction in the capabilty of producing trully novel Art (which AIs can't do since their models make the initial random noise which is the basis of their output coalesce towards what's statistically closest to that which they have been fed, which by definition must already exist), then a reversion in the use of Generative AI would be an undoubtoubly good thing, otherwise it really depends on the Economic impact of the job losses due to Generative AI versus the gains in speed of production (though with a lowering in quality) of using Generative AI.
As for the rest, yeah, there is a ton of what's called "AI" - let's just use the old terms for it, Machine Learning, aka ML - which does create Economic value which wasn't being created before and if it was lost then there would be a loss of overall Economic value.
This latest kind has by far received the least amount of investment money in this bubble, so this bubble popping and reversing would have very little Economic impact via this pathway.
Anyway, my point is that, as time went by and as the limits and capabilities of the Technology were discovered, it turned out most of the investment money into "AI" went into malinvestments which destroy overall Economic value (but can make a few people extremelly rich, hence why the money went there since they each economic actor only care about themselves, not about the whole Economy), and structural economic shifts to support said mainvestment just spread those economic malallocation further into the broader Economy. This "AI" economic tissue is akin to a cancer in an otherwise healthy body - a self-propagating bad mutation that extracts nourisment from the rest of the body, expanding to extract ever more nourisment whilst destroying the healthy parts - so excising it in it's totallity is not just a good thing but actually a requirement to restore a healthy Economy, and we should not fall into the trap of simplistically thinking of it as "part of the body" when in reality it's a previously healthy part of the body which underwent a carcinogenic mutation that turned it into a parasite that will eventually kill its host.
In simple terms, the harm and hence the damage is already done, and stopping and excissing it from the Economy now would not be the "harming of the Economy" (because that's already done), it would actually be the damage control necessary to save the Economy from suffocating due to excessive malallocation of resources.
How do I benefit from a massive corporation making a lot of money for shareholders when I don't have any shares and don't work there because they already replaced all of us with AI?
You won't, that's my point. The shitty system that we live in means that when a ton of companies crash at the same time ordinary people feel the negative economic effects. This what I meant as fallout.
How? What positive effect were they providing in the first place that I would lose.
They weren't providing any. But the share price on the US stock exchange for most of the big AI companies have been inflated through shoddy business practices, but they make up for an insane amount of the current trading of said stock market. When the bubble bursts and the market loses confidence the economy will contract and we hit recession or even worse depression. Much like the dot-com bubble or the financial crisis.
Layoffs. Massive layoffs. C suite will get massive bonuses to celebrate.
Would you like fewer employees? Dumber employees? More expensive employees? Try AI. It will bleed you dry and destroy your business. Available everywhere today!
Please please PLEASE help me FIRE my Employees!
-Billionaires who NEED to NOT be Taxed because they MAKE Jobs!
So AI is losing money hand over fist, and the solution is for AI to lose even more money? Thank god our economy is on a sound footing.
“We need to see the pricing for AI come down,” Arora said.
As a CEO, you really should know that prices never come down. Prices only go up. Gotta generate that shareholder value.
Prices for services like this do not ever decrease. By 2028, costs will likely be double what they are now.
More than that. At the current discount prices, LLM providers bleed money like mad. They will run out of cash soon if they don't raise the prices to profitable levels. Just double won't cover it.
Which will lead to most customers dropping the LLM services stat. Which in turn will force the LLM providers to up their prices even more...
When my last employer opened up a publicly available chatbot they asked us not to test it because of how much it costs. It was a sales bot, we had recently had the company decimated by redundancies.
A common act of protest was to abuse the public chatbot from a non work device.
The last 50 years has shown that the new model is to get a sector of the economy or consumers dependent on a service then once the competition has faded away, jack up the price. Competition either never shows up or collaborates on pricing.
Depends entirely on whether or not Moore's law has hit a physical limit yet.
If we're still doubling computational every X years for the same cost, we'd be able to see same models cheaper. If we're however finally hitting physical limits, then yeah. It won't be cheaper and we can't just smash more context in it like there's no tomorrow.
Then the most likely breakthrough would have to be DNA storage (tri point storage, rather than bi point storage) and then we'd need fast read/write to DNA storage. As that'd theoretically allow for more context for the LLMs, context is the biggest bottleneck
Not even once did he think whether it was useful or not. We really need to replace CEOs, managers, and the whole C-suite with LLMs, that would be useful for once! And at $20/month, it's a bargain and you can redistribute the money to people who really work.
He sort of did with executive double speak: he said two things by implying the cost wasn't worth it. To justify the poor quality the price has to drop by 90%. That's shareholder speak for "this is not profitable and not worth the debt exposure".
Nah. Ai also trained on ceo actions, so I'd say it would result in eviler mecha ceo driving the ship drunker than usual
We don't train then on ceo actions. We train them to run a company for the betterment of society, it's people, the employees and its survival.
That requires making decisions, which LLMs can't do. They can only emulate decision-making
I thin you overestimate the value of CEOs.
Not really, we're talking about running companies for the betterment of society, which CEOs don't do, so it's not implied that they would be capable of the thing that I mentioned would be required to do that
Also, they do make decisions. Almost exclusively bad decisions, because they're out of touch with the real world and the lived experiences of ordinary people, and they don't care about right and wrong, and their only concern is "line go up." But even bad decisions are still decisions.
LLMs literally cannot make decisions, good or bad, because that's fundamentally not a capability it possesses. If you ask it to make a decision it will hallucinate something based on any irrelevant details it finds in its context window. It can't perform cognitive reasoning, or weigh pros and cons, or consider secondary and tertiary effects.
Ha! Wait until they charge the actual costs and profits! You wish you had not kicked out you most valuable employees...
AI: "The best I can do is offer you a substandard "worker", but it'll cost you more."
"We are dangerously close to needing to consider the idea of near record profits as opposed to record profits... This is like 9/11 + that flood in the bible + when your Wi-Fi doesn't fully go out, but it's on/off again all day and it's like torture, you know?"
He thinks AI is expensive NOW? We are in the trial period. Wait until the whole thing pops and then tell us how expensive it is.
Get fucked. Globalprotect is a shit product that is about as reliable as a 40 year old renault and it being slop coded explains a lot.
Yet those prices can only go up. My rough estimate is about 10x higher to be profitable.
"The whole point was to increase profits by firing all these disgusting humans around here, not just pay the same money to you."
So when is the us gov gonna start subsidizing AI with your tax dollars to keep this bubble just on the solid side of bursting?
They can't.

How many CEOs have thought about how difficult it is to strike when workers have a family to feed?
Now think about how easy it is to scale down AI labour?
Also, you gave all your corporate secrets away to the psychopathic class. They will use this for insider trading, hostile M&A corporate intelligence for competitors, American empire building and hostile actions against foreign countries.
😝🤣😂😹
That's just the dumbest and most disconnected-from-reality take on AI I've heard in a while. Fire your CEO there is no way he's worth his salary.
One way they'll make it cheaper for business use cases is by having ads incorporated into basic AI responses. User: "Here is a picture of some rotting wood in my shed. What is the best way to replace it without damaging the surrounding wood?" Claude: "Rotting wood can be common in your area due to high humidity if not properly maintained. Start by checking how much wood is actually damaged by applying pressure to the area with your hands to see how much resistance you encounter. If only part of a board is rotten you'll need to replace the entire board. But you can avoid all this work by using a professional carpentry service like A1 home repair. They're in your area and open now. Their number is 407-322-9112. Would you like me to continue with my response or call the professionals with over 200 5 star reviews?
Wait until investors realize that executives can be replaces with AI also.
That’s just not gonna happen. Inference is incredibly expensive, and training is even more expensive. There’s not a lot that can be done to change that.
Just use smaller models, my guy. Humans are so easy to replace, a 30b model running on an old RTX 3080 ought to do it, right?
30B models are no featherweight at all and an old 3080 won't be able to lift them due to insufficient RAM and even if it did, low bandwidth would make it so you'd only be able generate a few dozens of tokens per minute.
This message alone would take like 2 minutes to generate.
A quantized 30b model will run on a 3080. I know because I’ve done it. But yeah, it needs to be quantized really small.
Edit: wait, no, you’re right, I’m thinking of a 3090. I’ve done it on a 3090, not a 3080. Quantized to 4 bits. And yes, it’s very slow. That’s enough to replace a human, right?
But the fun thing about the scenario of open source/light weight models dominating. Is the fact all of the leading LLM companies would go belly up, it's the biggest bear take on Anthropic, OpenAi and the likes - They need to have people use the biggest, most expensive models they can't run themselves, that's their entire claim to trillion
all 85 comments