This makes sense. However there is an issue where many people conflate "consciousness" with the ability to make novel insights, think genuinely, etc. They use this to claim that since AI is not conscious, AI could never actually "think" and is instead just always a regurgitation of its training data.
It is a natural human hubris to make our abilities seem unquantifiable and mysterious, but all the useful things the human brain does are just finding patterns in data, running lossy simulations, and estimating on abstraction, all things it is theoretically possible for any non-conscious machine to do.
To the best of my knowledge, there is not an agreed and actionable definition of consciousness, and any attempt to make one comfortably fails to cleanly divide humans from machines.
It's more of a vibey term, and as such it is genuinely very difficult (perhaps impossible even) to concretely determine whether an LLM possesses consciousness. LLMs successfully express a lot of consciousness-like traits.
At some point you have to ask the question: does it even matter? If an LLM can sufficiently mimic consciousness, isn't that sufficient for us to treat it as conscious, even if it is in-fact not conscious (especially because we don't actually know)?
As I've mentioned somewhere else already and you pointed out; the issue is that consciousness is a loose term and and it's a semantical issue that should be resolved first.
It won't be, because the murkiness is beneficial to the corpus, amongst other things.
And to your point about the appearance of "consciousness" being enough imho Chiang explains fairly well why it's not.
- Are current LLMs conscious?
- Is it possible that future versions of LLMs with similar architectures could be conscious?
- Can any AI be conscious?
I'd assign probabilities of around 0.1, 0.2, and 0.9. My completely ignorant take is that we probably need something more "dynamic" than a bunch of transformer layers in order to produce consciousness, but I wouldn't be shocked to be mistaken.
I assign probabilities of zero to all 3. Computer program being conscious leads to ridiculous and obviously false conclusions (think about a person running a program using pen and paper for memory).
If it was true, you can create extreme pain by running a program. You can run the program by simulating a CPU, using pen and paper for memory. So you're essentially claiming that some simulated being is in pain because there are some 1s and 0s on paper. In fact, you can decide to use an arbitrary encoding of the memory, so a sufficiently long sequence of 0s written on paper corresponds to a simulated being feeling pain in some encoding. That is clearly nonsense.
Time scale matters a lot in how we as humans perceive things like agency. Plants grow too slow for us to see any intent, but when you speed up a time lapse, suddenly it looks like plants reach for sunlight and vines for supports. Now, that may be projection on our part, but it may not be.
> The crux if it is that if you ever break from "the universe can be fully expressed mathematically", you are stuck in the mud of supernatural beliefs.
That's just not true. I've broken from "science can explain everything" and there's no mud. All of my beliefs are backed with careful reasoning. If there's an unknown, I don't fill it with random garbage.
Isn't pain just a manifestation of a bunch of chemical and electrical signals in the brain and body? It's not "clearly nonsense" to me that you could cause pain by writing a sufficiently long sequence of 0's - for it to be obviously wrong, you'd have to have some understanding of where consciousness comes from.
If you don't understand that, how can you assert that it doesn't come from mathematical relationships?
What specifically you mean by "manifestation of a bunch of chemical and electrical signals"? The brain is a system of physical particles. Why do I feel pain when the system is in state X but feel happy when it's in state Y? Physics can't explain that.
Do you seriously think that there's any chance that writing a lot of zeros on a piece of paper will create a feeling of pain in some conscious being?
This seems like a logical error. I don't understand how an internal combustion engine works, but I know it doesn't come from goblins jumping up and down inside.
The fact that you know it literally means you understand, at least to some extent, how an internal combustion engine works (i.e. it is powered somehow by combustion, and jumping goblins are not combusting generally).
If you would have zero knowledge about ICEs, how would you know?
Once or twice I've experienced extreme pain, and it was downstream of a bright light shining on a wet rock for millions of years.
I try to imagine myself long ago, on the outside looking in, with someone explaining to me that extreme pain, wondrous art, hunger, triumph, and despair would all unfold in due time where the rocks were wet and the lights bright enough.
I think the problem with this argument is that it's too inclusive. Is the bacteria that's adapted to an antibiotic conscious? It's showing intelligence right? I think if you're going to say something is potentially conscious, for me to take the argument seriously at least, there needs to be some plausible mechanism. I just don't see one for LLMs.
maybe the bacteria are conscious. How sure are we that they're not?
The only strong argument I have against it is the anthropic principle -- there are billions of times more bacteria than humans, so it's overwhelmingly unlikely that I'd be a human rather than a bacteria.
I have yet to be convinced that LLMs can produce definitive knowledge that is not a result of combining previous information. Humans can (if they can't then science basically collapses epistemelogically, see: philosophical skepticism), but I see no evidence of LLMs doing it. And from the number of truly new ideas and concepts delivered by LLMs (exactly zero), I think it's reasonable to just treat them as induction machines for now, but to treat anything they "know" as a Gettier case.
I would like to push back on the idea that humans can provide definitive knowledge that is not a result of combining previous information.
Pretty much every single idea in science can be traced back to some combination of earlier ideas, and as they get earlier / simpler, they can be related back to some combination of direct observations.
It's not clear to me at all that our entire body of scientific knowledge can't be simply recreated by "combining results of observations + previous information". And LLMs can perform observations in addition to combine previous information, which in my estimate is genuinely sufficient for them to plausibly be able to rebuild all of science.
How can you say for sure that every effect must have a cause? How can you be any more sure about that there can't be an effect without a cause than the old believe that black swans couldn't exist that was so universally believed that it became a cliche ("rara avis in terris nigroque simillima cygno")?
If every piece of knowledge is created by combining previous ones, then no true grounded knowledge is possible. Every bit of knowledge is grounded by another bit of knowledge which is itself grounded on, ..., etc, etc.
Your comment is conflating a few things here, the philosophical problem of induction is one I am very familiar with, and it's also unrelated to what we are talking about.
"If every piece of knowledge is created by combining previous ones, then no true grounded knowledge is possible" - this is not true, and has a very simple solution. What you are calling "grounded" knowledge comes simply from observation.
When the very first human saw that an apple falls of a tree, they have acquired what you call "grounded" knowledge. They can then combine that knowledge with the grounded knowledge that other things fall from trees to start reasoning about why things might fall from trees.
The problem of induction is different. It says, we know that things fall from trees because things have always fallen from trees, but how can be certain that what we call the laws of physics will not arbitrarily change on us?
And of course, it's a famous problem because it doesn't have a satisfying answer. The answer is basically "well everything we know is wrong if induction is wrong, so we will pretend induction is not wrong and hope for the best". And at least so far, that approach has seemed to work (heh heh heh).
Both of those are the same. Because the same shortcoming you mentioned (how can we be certain they won't arbitrarily change on us) can be overcome by making a compelling argument for synthetic a priori knowledge. Something (imo) put to bed pretty well a few centuries ago and only revisited because of the, uh, tedious nature of analytics, whose tradition evolved out of fancylad British boys who were all empiricism pervents.
It is a huge unspoken reality how much one's physical appearance affects the way they are treated, their life outcomes, and ultimate success in social/romantic relationships. Hair transplants, leg lengthening, plastic surgery, etc. will all explode over the next decade as AI erodes humans' ability to be successful via their industry and intellect.
The irony is that none of this stuff actually works as intended. Plastic surgery is obvious. Lip injections are obvious. Leg lengthening, I mean have you seen the proportions after?? Hair transplants too. If the most wealthy people getting these procedures look botched, what hope does anyone have really. Also, are we acting like steve jobs wasn't still a handsome man with his grey thinning head?
It doesn't come from getting legitimate validation from others. It comes from one's own fear of aging and their own mortality. Sorry, but we all shrivel up like a raisin by the end. Trying to beat that back with these means just seems so futile. Spend that cash on therapy instead to tackle your body dysmorphia.
Survivorship bias, you only notice the plastic surgery that isn't good. Most of the time it's invisible, your brain doesn't process the individual change, you just get the sense that the person looks better/less tired/more put together.
>Lip injections are obvious
Same phenomenon, but even if they are obvious some people like that aesthetic, in the same way that dyed hair/painted nails are obvious, but that's the point
>Leg lengthening, I mean have you seen the proportions after
For some men it is far better to be 6' with wonky proportions than 5'7" with perfect proportions. There is far more hate directed towards short men than men with long legs.
>Hair transplants too. I mean are we acting like steve jobs wasn't still a handsome man with his grey thinning head?
Not everyone is as handsome as Steve Jobs. If you have a handsome face you can get away with balding, if not then its a further infliction on how people percieve you.
>Sorry, but we all shrivel up like a raisin by the end
If we all die after 80 or so years then what's the point of doing anything? Why get a job, why put any effort into personal grooming?
> Survivorship bias, you only notice the plastic surgery that isn't good. Most of the time it's invisible, your brain doesn't process the individual change
So you're telling me all these Hollywood stars having infinite money, access to the best surgeons and are literally paid to look good get butchered on purpose?
Yes, often times it's some specific surgeon that overpromises, messes up, then subsequent surgeries which attempt to patch the issues end up failing. There are many celebrity surgeons who are good at marketing but bad at execution, and many celebs themselves who can't discern a good or bad job.
Everyone looks great if they live healthily and groom themselves. No one is actually an ugly duckling.
>If we all die after 80 or so years then what's the point of doing anything? Why get a job, why put any effort into personal grooming?
Because you want to be fit and healthy? All these surgeries are orthogonal to that. Every surgery is risky even medically necessary ones. You shower and groom yourself to prevent skin issues. You work out to simulate the hunter gatherer lifestyle your body is adapted to in the modern society which does not sufficiently pressure these adaptions. This stuff is your oil change and tire rotation. It is maintenance really. These surgeries are not maintenance however. It is like ricing the civic while it burns oil and the transmission makes scary sounds.
> Everyone looks great if they live healthily and groom themselves.
Not true for everyone. Simply living healthily and grooming, you are still limited by the ceiling imposed by your body. Sometimes it just doesn't live up to the aesthetic demands of the human psyche, used to a superstimulus of attractiveness as the norm.
>Every surgery is risky even medically necessary ones
Most cosmetic surgeries carry very little risk, your overall risk from even being a moderate drinker, commuting to work in a car, being mildly overweight are far higher in aggregate.
My main point is, the human modern social world isn't a perfect arbiter of reward based on whether you are doing all the "right things", by being generally healthy.
Sometimes people get a disease, like cancer, and nothing our body evolved to do can help, but human-invented therapies can actually help. Likewise, you can be perfectly healthy and nevertheless start balding, and no amount of generally being healthy is going to fix it, and no amount of generally being healthy will mitigate the minor albeit real social cost of it.
An exception isn't the rule. However Donald Trump by NY billionare real estate executive standards is fairly good looking, is very tall and has no major deformities. I doubt Trump would have gotten as far as he did if he were below average height, balding and didn't have expensive veneers.
Sure he would have. Plenty of people at the top of their industry are short, bald, and not conventionally attractive. People only have a good 10-15 years between looking 16 and showing signs of aging, sometimes much less. Such a small percent of one's life. But, people are not rational or logical but highly emotional. They see a grey hair and they feel bad. A wrinkle and they think it is the end for them. Sorry, you are becoming a high mileage car. This is just what happens. You got a lot more in you to go though.
I agree, but admitting the shallowness of the general human populace seems to be a moderate social taboo. You can get away with contextualizing it in a way that equivocates the nature of the phenomenon, but addressing it too directly seems to get pushback.
Saying
>I'm getting a hair transplant because I want to feel better about myself
Is received much more favorably that
>I'm getting a hair transplant because people are mean to bald people and I would downgraded a few points in implicit social status and general treatment if I were to exhibit male-pattern baldness.
The human brain responds positively to a full head of hair, and less positively to balding. I'm not sure why, as it seems there is no huge evolutionary advantage to hair.
However the way the brain responds to it is incontestable, and no amount of body positivity will change the impact it has, thus for an individual it makes much more sense to pay a few thousand to conform to beauty standards than hope that everyone you meet will turn off their primate brain when interacting with you.
This is all contingent on AI forays into mathematics being slop and low quality. However it's clear that recent AI models are capable of genuine mathematical achievements which surpass the frontier of what humans are able to accomplish (wrt the unit distance Erdos problem).
The issue is, how is a group of intellectuals, whose identity derives from their ability to do something rare, useful, and requires many years to get good at, react when a machine can produce all of their useful output nearly automatically, can verify its own outputs, and is getting better exponentially? It is the complete annihilation of one's sense of value and purpose when the binding element to your culture is commodified.
I think there will be a lot of arguments trying to claim that the point of mathematics is curiosity, or that there is always some ineffable human element that AI can't replicate, but I fail to see how somehow these wishy-washy human centered values somehow mean anything compared to the amoral pursuit of mathematical truth, which has nothing to do with humans.
It's just that we humans happened to be the only beings in the universe good at math until ~2025. Now there is another species which can do many of the things we do, and it is not bound by the size of the human brain, our short term memories, or the architectural limits of biological computation. To imagine that humans would retain supremacy in this very un-human like discipline seems like wishful thinking.
> This is all contingent on AI forays into mathematics being slop and low quality
It's literally a set of recommendations for researchers on how to use AI to advance the field and prevent slop from overwhelming the people who might do anything with the research produced.
For people who are so eager to declare that everyone else is just having an existential crisis because "your culture is commodified", AI people are getting awfully defensive about this document.
Another example of American Derangement Syndrome, where socially uncouth but generally well-meaning Americans (in the aggregate) are made out to be some sort of existential scourge. Many Europeans have this due to Americans seeming to have much more money than their taste would imply.
What is the advantage to not keeping them a secret? The populist movement against AI is growing rapidly, and is supported by bunk science which affirms people's pre-existing biases (like the idea that data centers suck up all the water in a community, or raise the ambient temperature by a single-double digit number of degrees F).
AI is a genuine source of economic growth. I can understand wanting to curtail it, but in return you are getting fewer jobs, less economic growth, more money to other countries who don't allow protesting or even complaining about data centers, etc.
The US has deindustrialized and since the rise of NIMBY politics hasn't built new industry in any scale at all. The first freeway revolts happened in 1959. In the meantime, basically everyone has agreed that industrial growth and manufacturing as an abstract concept is Good (TM) but definitely not in my backyard. For a long time these kinds of things were just being built in neighborhoods where the residents were too poor/politically disconnected to organize, but eventually that too became difficult.
Since automation became big in industrial processes, most industrial development has been labor-poor (few workers) but continues to be land intensive. That means while industry might generate tax revenue, it doesn't have a coalition of labor advocates willing to champion the capacity because not many jobs are created. So we find ourselves in 2026 with an inability to actually build new industrial capacity in any form. The anti-tech crowd is angry at the data centers, but the same exact thing is happening when it comes to permitting new power generation and transmission lines. In fact many of the concerns related to data center power usage could be allayed if we had more power generation but nobody wants power generation in their backyard.
For decades now the US has been dancing around the idea that there is no by-right way to build anything anymore, so building any new large structure becomes a collective action problem that just ends up failing. Even when things get built, costs are massive. This has even affected things that the US is ostensibly really good at making such as highways, as evident in the recent Texas highway expansions ongoing.
I think it's good to remind people that the industrial revolution was very close to never happening.
This is from Dud Dudley writing in 1665, whose own ventures to manufacture steel en masse before Abraham Darby succeeded.
> "I have been opposed by many adversaries, as by wood colliers, mine owners, and others who, being poor men, did, by misguided advice, throw down and destroy two of my furnaces and my works, and caused much of my pigs and bar iron to be carried away."
There were plenty of examples through history of "near-misses" where establishment land/wealth holders suppressed nascent steel industries. It was almost an accidental series of coincidences that the industrial revolution happened - the Glorious Revolution in England and Abraham Darby's secret financing network.
It should also be remembered that while the industrial revolution netted humanity enormous wealth and eventually a higher average standard of living, it also kinda sucked for the generations of working class living through it, prior to labor reform. Millions of people lived entire lives where the industrial revolution was nothing but bad for them and never saw the upside. So anybody opposing a new industrial revolution is not necessarily acting out of irrationality.
Kind of. I think the history is interesting here and complex.
One of the hard things to grasp is that the industrial revolution was preceded by an environmental collapse. Part of the reason there was a switch to coal (despite being seen as inferior to wood at the time) was massive depletion of wood in England and the high cost of importing not just timber but even just firewood.
Add this in to the enormously expensive wars England was fighting all through this period and stressed everything from labor and food supplies (which also triggered demand for steel and copper and brass) The industrial revolution happened against a backdrop of national crisis so it's hard to know what was being caused by the revolution and what the revolution was helping paper over.
And on top of this, when Engels and Marx wrote about the squalor and desperation of their time (which was very real), nearly a hundred years had passed and something much different was happening. Massive amounts of peasantry were being dispossessed of lands and forced into urban slums. Cities grew something like 10x in a single generation. This wasn't really the fault of the industrial revolution but because of really bad policy.
Now obviously sometimes new developments really are bad, or they're being targeted at terrible spots -- you don't want a very polluting factory right next to a residential neighborhood.
But we've overcorrected drastically. The rules should be sensible and out in the open; planning committees should only be checking whether companies and governments have followed the law as written, not listening to every single possible objection from neighborhood residents about a new apartment complex affecting street parking or creating shadows.
As I pointed out in another comment in this thread, a "neighborhood group" that's using environmental rules to block a low-cost, employee-owned grocery store being added to a site that was already basically a grocery store (Sam's Club) before, is an insane weaponization of the rules. Nobody who was writing these laws ever intended environmental protections to be misused like this.
This is an excellent analysis. The NIMBYism is pretty self-evident I think but I never thought about the disconnect automation may have caused between industry and advocates for industry.
>basically everyone has agreed that industrial growth and manufacturing as an abstract concept is Good (TM) but definitely not in my backyard. For a long time these kinds of things were just being built in neighborhoods where the residents were too poor/politically disconnected to organize, but eventually that too became difficult.
These NIMBY jerks have redefined the entire country to be their back yard
The NIMBYs were pissy that the poors were allowing industrial development said "those poor people are stupid and wrong for accepting the industry" and the legislated hoops so that they essentially have a "say" in what these poorer communities can allow.
They usually leverage environmental laws to do this stuff because the richer areas are already developed and paved to high heaven. They aren't doing greenfield development and the laws are construed to basically punish/prevent greenfield development and allow brownfield development and re-development. Of course, residential gets exemptions to all sorts of stuff so their ADUs and the kind of development they do are mostly unaffected but the industrial stuff out in the sticks is all but prevented.
I see this in my own city that's over an hour drive from the rich places. We want to allow manufacturing but the state will take our grant money if we don't have and enforce a ton of rules and process that make that prohibitive so basically no industry can afford to create a facility except the biggest of BigCos.
You're conflating AI and AI Data centres. What kind of "jobs" does AI data centre creates? More jobs for factory workers who assemble video cards to be shipped and installed in a rack?
The downstream portion isn't proven, and more importantly, even if it was, it's not localized to where the data center is built (and its effect on the local population).
How many technicians? And from where? Its unclear to me how often this stuff actually needs to be done. And how resilient the overall system is to failure. Is it imperative to swap out one point of failure immediately or can you let them batch up and send Joe from California out qaurterly?
Technicians are constantly behind in work. Many data centers are pulling 24/7/365 shifts to keep up with demand. Larger data centers have hundreds of full time employees.
Yes many servers are left in fail over states for long periods of time, but that can only be done because new capacity is actively being deployed to make up for that fail over. Modern data centers are far too big for a single person to be repairing things every once in a while. Stuff is breaking every hour of every day
Thanks for the info. So the technicians are there on site daily? dozens? hundreds?
Im not too surprised there is quite a bit to replace. Sure the failure rate should be low but the scale is massive. Its interesting to get an order of magnitude though. Every hour is more frequent than I thought.
I have let's say 17 data centers being built in my town. 1700 jobs is not enough according to the people in opposition. The real number is likely higher, and my town is under 30k people.
I worked for a company that ran a pretty big, high-security data center--75,000 square feet serving thousands of customers. There was a team of 6-8 technicians and a couple dozen security guards. None of them were particularly great or high-paying jobs.
Regarding water usage, in general data centers do not use more water than other types of heavy manufacturing
> The Georgia data center is only using ~2% of the county’s water. For comparison, a pharmaceutical manufacturing plant is using ~4% of the county’s water. A construction plant for Rivian cars is using about the same amount of water as Meta’s data center. The data center is functioning like any other normal industry in the county.
Regarding data centers increasing ambient temp, the paper is simply measuring the surface temperature of the buildings, going against the claim that a data center, merely by its presence in a community, raises the ambient temperature by a few degrees or more
>Regarding water usage, in general data centers do not use more water than other types of heavy manufacturing
Are the claims really that "Data centers use more water than other types of heavy manufacturing"? I dont think so.
Even if thats true, that doesn't mean they cant have a disastrous effect on the local water supply. This isnt a good rebuttal.
Frankly I tend to think the anti-datacenter crowd is overreacting. But I don't think you've addressed the real criticisms being levied.
In some passing research I saw the datacenters do continuously consume water (its not a one time cost like some claims I've read). And smaller size ones may use water equivalent to around 1000 households, and larger ones may consume closer to the equivalent of around 20,000 households. Evidently the massive one in Utah will at least double the state's entire consumption of water.
Can all of these places handle it?
I dont know. But that's the question, not if other types of heavy manufacturing have higher demands. And frankly it's inevitable that at least some locations cannot handle it. Which doens't mean you should be anti-datacenter in general. It means you can't just blanket dismiss the water concern for all locations.
Data centers do use water, and a lot of it, but the claims being made are hyperbolic and not squared with reality.
One criticism I often see is that data centers somehow pollute the local water supply. Data centers use water in a closed loop, their impact on local water quality is negligible. Industrial manufacturing and even agriculture have a far greater deleterious effect.
> The EPA’s national assessments repeatedly identify agriculture as the leading source of impairment for rivers and streams due to nutrient and sediment runoff, with continued nitrogen and phosphorus problems that affect drinking water and coastal ecosystems.
The thing is, AI data centers bring in far more tax revenue than other water-guzzling domains (like golf courses), but use less water overall. Extreme panic over specifically their water use is disproportionate
> How much of this will be AI? Almost all this growth will be driven by AI, but because AI is only 20% of data center power use, its growth will have to be huge to triple total power usage. One forecast says AI energy use in America will be multiplied by 10 by 2030. Because water use is proportionate to energy use, we can multiply AI’s water use by 10 as well.
> So in 2030, AI in data centers specifically will be using 0.08% of America’s freshwater. This means it will rise to the level of 5% of America’s current water used on golf courses, or 5% of U.S. steel production, or be about 173 square miles of irrigated corn farms.
> The average American’s consumptive lifestyle freshwater footprint is 422 gallons per day. This means that in 2023, AI data centers used as much water as the lifestyles of 25,000 Americans, 0.007% of the population. By 2030, they might use as much as the lifestyles of 250,000 Americans, 0.07% of the population. Not nothing, but 250,000 people over 5 years is just 4% of America’s current rate of population growth. If you found out that immigration plus new births in America would increase by 4% of its current rate, would you first thought be “We can’t afford that, it’s way too much water”?
"Closed loop" doesn't mean no net water use after filling. There are leaks, and the water in the system needs to be processed for reuse, and that processing needs clean water.
Even if there is no next water use, "closed loop" refers to cooling the data center proper, and excludes the water for the (primarily) thermoelectric power plants which power those data centers - a power load which is higher due to using closed loop cooling instead of evaporative cooling.
Given that many of these are the same companies which once promised net-zero CO2 emissions by 2030, you'll excuse me if I insist on full information about the total environmental impact and tearing up all of the NDAs they require from local governments.
How many liters per kilowatt-hour does each site use? How much CO2, NOx, and particulates are produced? What are the power sources? Why are EPA waivers needed and appropriate?
This should ideally include the supply chain - those GPUs need a lot of very pure water, and 83.2% of Taiwan's power and almost 60% of South Korea's comes from fossil fuels.
> their impact on local water quality is negligible.
So there should be absolutely no issues in publishing all this information, right?
> far more tax revenue than other water-guzzling domains (like golf courses),
Which people already complain about because they use too much water, and often exist only because rich people got special arrangements. For some examples of the antipathy for the Santa Fe Country Club and golf courses in general, see https://www.reddit.com/r/SantaFe/comments/w9g4ak/the_city_of... .
But even the Santa Fe Country Club case highlights how tax revenue is only part of the total economic benefit. For example, they were allotted 700,000 gallons of treated effluent per day, in exchange for public golf access with reasonable fees. While data centers typically used treated water, not treated effluent, and don't allow public access or activities.
For that matter, local birders visit the municipal course, Marty Sanchez Links, to see the birds using the water features and irrigation pond. Not a benefit a data center will offer.
From what I hear, surrounding residential prices go up around a golf course, and down around a data center, so looking at just a single entity's tax revenue isn't enough. To say nothing of the special tax deals the data centers insist on.
Under NDA, of course, which should be illegal for this sort of issue.
> AI in data centers specifically will be using 0.08% of America’s freshwater
Since you think these centers can be sited anywhere, why are these data centers being put in water constrained places like Utah, rather than water rich places like Michigan?
> The average American’s consumptive lifestyle freshwater footprint is 422 gallons per day.
Sante Feans use under 100 gallons per capita per day.
If you think the average American use is relevant, then put the data centers some place where there's water.
> If you found out that immigration plus new births in America would increase by 4% of its current rate, would you first thought be “We can’t afford that, it’s way too much water”?
Water use per capita has been decreasing over time due in part to mandated water-efficient fixtures and appliances, but also (at least in New Mexico) to changing practices like allowing xeriscaping in places which once mandated lawns, rain barrel and cistern rebates, mandated toilet retrofits, and water use awareness programs.
> This means that in 2023, AI data centers used as much water as the lifestyles of 25,000 Americans, 0.007% of the population. By 2030, they might use as much as the lifestyles of 250,000 Americans, 0.07% of the population.
A 10x increase in AI data center buildout between 2023 and 2030 seems unlikely, given the large number of AI data centers either in progress, or in the planning stages.
>One criticism I often see is that data centers somehow pollute the local water supply.
Yes, I think this one is completely misinformed. For example AOC held up some dirty water from a local resident's tap. Fine, that's bad. But it was a result of digging during the construction process and the fact that it was a data center was irrelevant. And the implication that it permanently ruined local's water supply was just wrong.
>The thing is, AI data centers bring in far more tax revenue than other water-guzzling domains (like golf courses), but use less water overall. Extreme panic over specifically their water use is disproportionate
I guess that's not surprising to hear. A lot of people against data centers are probably also against golf courses though. I think AI is valuable but a lot of opponents see it as a net negative. Not saying they are right - this is definitely a point against the anti datacenter crowd. But it is consistent from their perspective so I dont think this point will persuade them. Would need to attack the claim that AI is a net negative.
> > The Georgia data center is only using ~2% of the county’s water. For comparison, a pharmaceutical manufacturing plant is using ~4% of the county’s water. A construction plant for Rivian cars is using about the same amount of water as Meta’s data center. The data center is functioning like any other normal industry in the county.
How much employment and localized value/tax revenue is created by the pharma plant compared to the data center to offset the environmental effect?
If only they could spur even more economic growth with the planning and construction of more fabs for all the RAM it will need (assuming the level of growth that their S-1s are claiming)
> I can understand wanting to curtail it, but in return you are getting fewer jobs,
since AI -- as designed to be deployed -- is taking away jobs, then by enabling data centers you are in fact getting fewer jobs
yours is the same perverse argument that was used to build Walmart Supercenters (lower prices! they'll create jobs!) which literally destroyed the economy of many small towns and communities, and paved the way for every other big box retailer to do the same
so don't be surprised that people aren't eager to enable something that is designed to destroy their livelihood
By the time any research study is done on AI is published the models are already 0.5-1 generation ahead. Even this bullish outcome for AI models and their ability to perform useful work does not reflect how good they are now.
Why do you think that? All these rumors about compute constraint just seem like speculation and not based on any data or information. All they would need to do is increase their prices to free up compute capacity.
Mythos was announced a few month ago and has been actually demoed in many companies who have all reported its abilities, supporting the claims made by Anthropic. How is this in any way similar to the FSD situation?
You're conflating the protracted promises for full-self-driving with the current rollout of autonomous driving features in Teslas, a feature people are using today. I've driving with multiple people who use their Tesla self-driving and report its quality/accuracy, this isn't some overpromised future feature.
And I mean, it's not like Anthropic is a zero-product company that is only offering gated access to their only product, Opus 4.7/4.8 are very good and are driving billions in revenue. Anyone can use it and see how good it is, and it is clear that it is a very good model at many things. It is no huge leap to imagine that a model that is 10x bigger is also better at many of the tasks that Opus is good at.
They are gating the release because of cybersecurity/misuse concerns, which makes sense because
1. Existing models are already being used to find exploits and hack into systems
2. We don't know the effects of releasing a tool which can autonomously exploit systems, especially in a world driven by a "security through obscurity" philosophy. It makes sense to give a heads-up to patch up software that affects billions of users before releasing it.
Imagining that this delayed rollout is all a big marketing scheme, that they have gotten dozens of multi-national companies to play along, and that Anthropic is somehow now just patently being dishonest about something while they have every incentive to not be dishonest (especially when they are neck and neck with OpenAI and their relative success depends on verified claims about model abilities), is pure conspiratorial thinking and driving more by a motivated cynicism about AI companies rather than a reasoned examination of the claims being made.
Mythos was announced a few month ago and has been actually demoed in many companies who have all reported its abilities, supporting the claims made by Anthropic. How is this in any way similar to the FSD situation?
Exactly. Incredibly hard to understand what hard, non-headline-quoting, steel man arguments there are about how exactly the market will hiccup. And as if all of the AI companies somehow know this and are looking to IPO themselves out when anthropic revenue is growing > 10x per year for multiple years. Feels like a massive disconnect between “this will all implode” people and any real numbers.
All rallies do come to an end. The fact that we all don't know exactly what will cause this one to end is exactly part of the problem and 100% doesn't mean it won't happen. Usually some external shock spooks the market and a massive sell off happens.
So what could happen, any number of things. An obvious near term issue might be inflation increases dramatically in the US (on account of the oil shock), causing interest rates to increase - maybe dramatically - , which causes the stock market to retract. Also, the housing market is pretty much toast at the moment and an increase in interest rates might finish it off too causing a contraction there. So many ways things can break.
But honestly, I'll tell you after it happens and it will happen. Having lived through a few of these now when everyone tells you it's a sure thing and prices go up for ever you get an inkling you are near the pop.
> All rallies do come to an end. The fact that we all don't know exactly what will cause this one to end is exactly part of the problem and 100% doesn't mean it won't happen. Usually some external shock spooks the market and a massive sell off happens.
Yes sure, but that statement contains zero information -- why do you believe it will end in a time short enough for the "market bubble" comments to even make sense?
External shocks -- sure of course. Inflation problems in US -- absolutely, it's a ticking time bomb with a debt crisis looming. Housing market I don't really know anything about but I'll take your word for it.
But all of this has been true for awhile, and could have been stated with equal veracity over the course of the last 5 years at least. Your beliefs shape your actions; so why does this belief shape actions any differently than it has earlier?
> ut honestly, I'll tell you after it happens and it will happen. Having lived through a few of these now when everyone tells you it's a sure thing and prices go up for ever you get an inkling you are near the pop.
Again totally true, I have also lived through them and expect more. But "these companies are IPO'ing because they know the market will pop" is kind of the thing that I was trying to address. For all the signals of market danger, there are plenty of optimistic signals all over the data. Growth is pretty robust across all sectors today.
I think the original comment was before the market sneezes. So we all know it will sneeze and the companies want to get in on this business cycle. The longer they leave it the more likely they will miss it.
They need a price consumers can't stomach or are unwilling to pay, and without that the company is profitable but not able to justify investments. That's the argument.
I'm interested in your argument but who needs a price for what exactly? Like token costs are unsustainable argument? Or are you saying they need a stock price to keep their valuation high?
Token cost might be sustainable but still not profitable enough to make up for the costs of either training or the investments gifted away until it became profitable. They also might have no relevant moat that allows them to enshittify enough. But mainly, they are in "I need to kill whole industries to be worth it" tiers of investment.
> But mainly, they are in "I need to kill whole industries to be worth it" tiers of investment.
Yes agreed. Coding is a pretty big industry though in and of itself. Same with healthcare, legal, etc etc etc. Of course we have zero model today that can seriously kill an industry, but if you look at (1) how good things are today (insanely fast and rapid adoption) and (2) robust performance trends from many complimentary sources, it's kind of inevitable and I haven't really heard a coherent steel man argument for why "killing whole industries" is somehow a far-fetched idea.
> still not profitable enough to make up for the costs of either training or the investments gifted away until it became profitable.
Regardless of the weeds of the economics today, you have a clearly valuable asset that at the very least already a must-have for enterprise and will become even more essential over time. There is token economics that either already do or will make sense. You will have some sort of marginal cost + profit margin that things will stabilize at. You can pay a premium for high quality frontier models. "But it costs more in R&D to fund this!" ok but then token costs will increase. Why is this some sort of death knell?
>it's kind of inevitable and I haven't really heard a coherent steel man argument for why "killing whole industries" is somehow a far-fetched idea.
They don't only require "good enough to kill industries" (which is doubtful but certainly feasible), that's just step one. I think about it in terms of potential failure modes:
- if models don't reach worker-substitution levels, they fail
- if models reach that level, but it's too expensive to run and a worker's still cheaper, they fail
- if models reach that level, but the resulting tech is cheap enough to use, they fail (since open models can compete)
- If the models work but there's social rejection leading to regulation (due to mass unemployment for example), they fail
- if the models work but there are significant deal breakers (like a fundamental inability to keep them safeish to use) they fail.
So it's not really a single AI killer reason, it's more that the success case requires things to land in a very specific future where models work, and they're cheap enough, and expensive enough, and valuable enough, and exclusive enough, and safe enough, and...
Each "and" is a multiplier reducing their chances, and there's a ton or factors. Not imposible, but not where I'd put my money.
If 1B people consume via personally or their employer on average $100/mo of tokens or services that solves the problem right there and that is not inconceivable whatsoever and feels like an underestimate. Repayment doesn’t imply you have to build an end to end replacement of an industry.
This also negates this whole like “you have to completely replace a human” fallacy. Why do you need to replace a human? Why not just increase the value each human brings you?
The model of open weights has been around for awhile. You have frontier labs releasing them. They are powerful and capable and track yet lag frontier models. Without massive government subsidies from probably China who did something similar with manufacturing I don’t get the idea of OSS somehow toppling the entire industry of frontier models. How would this happen? Did demand for frontier models drop after GPT-4? GPT-5? Because the cost of GPT-4 perf is maybe 100x cheaper today. You can always pay a premium for a better model (more data spend with proprietary data sources, more compute for training, more thinking budgets), and in the end, things will either go to a monopoly or prices will stabilize around the marginal costs.
Also safeguards are always important. They are thorny fundamental horrific problems and I can tell you there will be a hellscape of pain as people figure out the trivial ways to do bad things with some of the worst security practices we’ve seen. But I don’t get why this is a dealbreaker, we are using these systems everywhere and in loads bearing environments, today, and reliability and hallucination rates continue to increase / decrease.
>If 1B people consume via personally or their employer on average $100/mo of tokens or services that solves the problem right there
0.15B is the full actively employed population of the US. 0.2 extra for the full EU, and we're barely making a third of the needed numbers, and this is assuming that the guy cleaning the street or the plumber will have a 100 bucks subscription just because - not to mention that a hundred bucks is 10-20% of median monthly income in many countries, even some EU ones.
You're asking for 3x Netflix subscribers at 4-10x Netflix price and with the whole world standardized at American prices (without American income).
Then add that this would very likely be a commodity market with competitors, so the billion, if existing, would be a market rather than a specific company's income.
one possibility is that some heavily indebted AI infrastructure company will be unable to meet its dept obligations, which will cause banks that become heavily exposed to AI-infrastructure related dept to tumble
I would guess fallout from the straight of hormuz closure. I don't think we have seen the full effects of that yet. Between climbing inflation, and a possible recession, we could definitely see the bubble finally pop.
It is a natural human hubris to make our abilities seem unquantifiable and mysterious, but all the useful things the human brain does are just finding patterns in data, running lossy simulations, and estimating on abstraction, all things it is theoretically possible for any non-conscious machine to do.
reply