It’s annoying to see AI executives attempt to spread fear and panic throughout society. We get it. They don’t value humans and are desperately planning to capture the market away from labor. The fact these executives have the audacity to boast about mass layoffs reveals how divorced from the consequences they think they are. It pisses me off! Am I alone in this?
> Without these chips, new data centers would not be built or would sit empty. Astronomical tech valuations, and in turn the entire stock market, could collapse.
My understanding is that data centres are already not getting built. It's one of the bellwethers that the AI bubble is currently in the process of collapsing, even if it's not immediately obvious from the outside.
Only a third of announced datacenters are actually in development, nearly 60% of the datacenters under development do not even have a plan as for how they'll get power.
edit: this isn't new by the way, a very large part of basically everything every AI company has announced has turned out to be nothing but hot air. From investments (remember that $100 billion Nvidia was gonna invest in OpenAI? Yeah that never happened), to predicted timelines until profitability which get pushed back a year every year to the outrageous claims of what the LLM's can do like create a teams clone except nobody can see it or the code and nobody gets to know if it actually works.
And then there's shit like Nvidia expecting more and more sales every year but it takes about 6 months to install 3 months worth of sales so where are all those GPU's and who is buying them?
AI is having a profoundly negative impact on our society. This article does a great job discussing AI’s fragile state and vulnerabilities that could send us into a downward spiral.
The issue isn't really about what AI can or can't do; it's about how people use it to affect the workplace. This matters more than anything. If everyone has to use AI for some reason then it becomes a nonsensical game of figuring out how to sell something people may not even want. As a teacher AI is creeping into online platforms whether they advertise it or not, so it's affecting what we teach and how. I'm 100% certain it's reading my students' answers and using them to improve its own instruction. But the question is about how one can have a business with or without AI and money dictates what you have to do. We're probably going to be required to do more work for less pay and definitely less certainty about where we'll have to be in one, five, or ten years. How can anyone plan their life ahead in that case? So then AI will probably be used to push people out of the workforce. Maybe you'll need someone for a few years, then you won't. Then it'll change and you need them again. But it'll be someone else. This isn't sustainable because businesses and tech people want people not just replaced but in this perpetual cycle of retraining and improvement, but this can't work outside of these industries. Not for a life worth living.
I suppose a panic is the expected opposite of the inebriated worship of AI we've seen from some in the past 5 years or so. But the reality will be somewhere in the middle. AI replacement of labor will occur, to a minor degree, allowing a small percentage of cash to be recouped. Some AI firms will go bust, to be shored up by bailouts or foreign purchase at POTD. Finance will be forced to trim some fat and the riskiest may fold up. The DOW may even dip below 40k briefly.
And by 2030 or so we'll have the next 'golden goose', be it AI hospitals or emergent agriculture or flying cars. Speculation is the system. It's why pensions have been replaced by 401ks, why every large-ticket purchase is about 'resale-ability', why our economic anxiety is burshed aside as moral failing.
The reason we shouldn't be afraid of AI is because economic value, real value, comes from labor. Our system might feed us 6 layers of icing and call it value, but we can taste the difference. The jobs that AI is primed to replace- call centers, PR people, financial advisors, data anlalysts, etc.- are not those that inspire or invigorate (See Graeber 2018). If this AI bubble goes more belly-up than belly-out, perhaps it will push our politicians to reorient what kinds of jobs really matter.
> Some AI firms will go bust, to be shored up by bailouts
None of them will get bailouts, there's nothing there to bail out because the underlying businessmodel is so bad. None of the companies are anywhere near important enough anyway, it's just venture capital throwing money into a furnace to chase growth for the sake of growth.
Remnants will be bought up by the big tech companies, for as much as they don't already own the IP through their investments anyway.
What bothers me most about that text is how quickly it moves from justified skepticism to a sweeping indictment. The fact that there is a lot of money, a lot of posturing, and a lot of political opportunism around AI does not by itself mean that every investment is pointless or that the entire sector rests on a conscious disregard for reality.
It would be stronger if the author moralized less and sorted things out more. For example:
what in AI is already demonstrably useful,
what is just a marketing label,
which expenditures are a reasonable infrastructure bet,
and which ones already look like a late-stage manic competition.
That is exactly what the text is missing. Instead, we get phrases like “bent the knee,” “polycrisis,” “ungodly amounts of capital,” and other language designed more to provoke the reader than to inform them.
Skepticism toward the AI boom is legitimate. But if it wants to be convincing, it has to rest on measurable things: return on investment, energy readiness, supplier concentration, and real revenues from deployment. Without that, it is more atmosphere than analysis.
I don't think we're going to get balanced views until it's all over - too many unknowns, too much temptation to project your own hopes and fears onto the crystal ball. I see two "doomer" camps - first the Luddite camp in wider society, but there's also a thread of "doomer hype" from inside the industry that goes something like "it's gonna replace 97% of jobs in the next three years, you better invest heavily so you don't fall behind!" That one pisses me off deeply, and it's very "mask off" of the business leaders who are peddling it.
Having worked with them quite a bit over the previous year, I've come to the conclusion they're best described as translation engines. They translate natural language to natural language, natural language to computer code, computer code to natural language, questions to answers, etc etc. If you can phrase your problem as a translation problem, that's probably where they're strongest - for example, they're by far the best thesaurus mankind has ever invented. I think this model makes sense given the way vector databases are structured, and explains why some limitations won't be overcome.
Most of the improvement now isn't going to be in the engines themselves, it's going to be in the UX we use to interact with them. I think there are still huge gains to be had there.
Mike-Banachek | a day ago
It’s annoying to see AI executives attempt to spread fear and panic throughout society. We get it. They don’t value humans and are desperately planning to capture the market away from labor. The fact these executives have the audacity to boast about mass layoffs reveals how divorced from the consequences they think they are. It pisses me off! Am I alone in this?
ours | a day ago
Mass capturing labor without offering an alternative is going to end badly. It's going to be bleak for a while.
Quouar | a day ago
> Without these chips, new data centers would not be built or would sit empty. Astronomical tech valuations, and in turn the entire stock market, could collapse.
My understanding is that data centres are already not getting built. It's one of the bellwethers that the AI bubble is currently in the process of collapsing, even if it's not immediately obvious from the outside.
Aethien | a day ago
Only a third of announced datacenters are actually in development, nearly 60% of the datacenters under development do not even have a plan as for how they'll get power.
edit: this isn't new by the way, a very large part of basically everything every AI company has announced has turned out to be nothing but hot air. From investments (remember that $100 billion Nvidia was gonna invest in OpenAI? Yeah that never happened), to predicted timelines until profitability which get pushed back a year every year to the outrageous claims of what the LLM's can do like create a teams clone except nobody can see it or the code and nobody gets to know if it actually works.
And then there's shit like Nvidia expecting more and more sales every year but it takes about 6 months to install 3 months worth of sales so where are all those GPU's and who is buying them?
[OP] hissy-elliott | a day ago
AI is having a profoundly negative impact on our society. This article does a great job discussing AI’s fragile state and vulnerabilities that could send us into a downward spiral.
pillbinge | a day ago
The issue isn't really about what AI can or can't do; it's about how people use it to affect the workplace. This matters more than anything. If everyone has to use AI for some reason then it becomes a nonsensical game of figuring out how to sell something people may not even want. As a teacher AI is creeping into online platforms whether they advertise it or not, so it's affecting what we teach and how. I'm 100% certain it's reading my students' answers and using them to improve its own instruction. But the question is about how one can have a business with or without AI and money dictates what you have to do. We're probably going to be required to do more work for less pay and definitely less certainty about where we'll have to be in one, five, or ten years. How can anyone plan their life ahead in that case? So then AI will probably be used to push people out of the workforce. Maybe you'll need someone for a few years, then you won't. Then it'll change and you need them again. But it'll be someone else. This isn't sustainable because businesses and tech people want people not just replaced but in this perpetual cycle of retraining and improvement, but this can't work outside of these industries. Not for a life worth living.
Cultured_Ignorance | a day ago
I suppose a panic is the expected opposite of the inebriated worship of AI we've seen from some in the past 5 years or so. But the reality will be somewhere in the middle. AI replacement of labor will occur, to a minor degree, allowing a small percentage of cash to be recouped. Some AI firms will go bust, to be shored up by bailouts or foreign purchase at POTD. Finance will be forced to trim some fat and the riskiest may fold up. The DOW may even dip below 40k briefly.
And by 2030 or so we'll have the next 'golden goose', be it AI hospitals or emergent agriculture or flying cars. Speculation is the system. It's why pensions have been replaced by 401ks, why every large-ticket purchase is about 'resale-ability', why our economic anxiety is burshed aside as moral failing.
The reason we shouldn't be afraid of AI is because economic value, real value, comes from labor. Our system might feed us 6 layers of icing and call it value, but we can taste the difference. The jobs that AI is primed to replace- call centers, PR people, financial advisors, data anlalysts, etc.- are not those that inspire or invigorate (See Graeber 2018). If this AI bubble goes more belly-up than belly-out, perhaps it will push our politicians to reorient what kinds of jobs really matter.
Aethien | a day ago
> Some AI firms will go bust, to be shored up by bailouts
None of them will get bailouts, there's nothing there to bail out because the underlying businessmodel is so bad. None of the companies are anywhere near important enough anyway, it's just venture capital throwing money into a furnace to chase growth for the sake of growth.
Remnants will be bought up by the big tech companies, for as much as they don't already own the IP through their investments anyway.
Optimal-Hunt-3269 | a day ago
I was with you until your last sentence.
tiger_overrider | 23 hours ago
What bothers me most about that text is how quickly it moves from justified skepticism to a sweeping indictment. The fact that there is a lot of money, a lot of posturing, and a lot of political opportunism around AI does not by itself mean that every investment is pointless or that the entire sector rests on a conscious disregard for reality.
It would be stronger if the author moralized less and sorted things out more. For example:
That is exactly what the text is missing. Instead, we get phrases like “bent the knee,” “polycrisis,” “ungodly amounts of capital,” and other language designed more to provoke the reader than to inform them.
Skepticism toward the AI boom is legitimate. But if it wants to be convincing, it has to rest on measurable things: return on investment, energy readiness, supplier concentration, and real revenues from deployment. Without that, it is more atmosphere than analysis.
PersistentBadger | 22 hours ago
I don't think we're going to get balanced views until it's all over - too many unknowns, too much temptation to project your own hopes and fears onto the crystal ball. I see two "doomer" camps - first the Luddite camp in wider society, but there's also a thread of "doomer hype" from inside the industry that goes something like "it's gonna replace 97% of jobs in the next three years, you better invest heavily so you don't fall behind!" That one pisses me off deeply, and it's very "mask off" of the business leaders who are peddling it.
Having worked with them quite a bit over the previous year, I've come to the conclusion they're best described as translation engines. They translate natural language to natural language, natural language to computer code, computer code to natural language, questions to answers, etc etc. If you can phrase your problem as a translation problem, that's probably where they're strongest - for example, they're by far the best thesaurus mankind has ever invented. I think this model makes sense given the way vector databases are structured, and explains why some limitations won't be overcome.
Most of the improvement now isn't going to be in the engines themselves, it's going to be in the UX we use to interact with them. I think there are still huge gains to be had there.
Last_Limit_Of_Endor | a day ago
Is there a back up of the article ? I clicked it and apparently the page is down for repair ?
FisherPrice | 17 hours ago
The irony of the audio version at the top of the article being generated by ElevenLabs is 🤌