Everything else is up around 3% YoY. And if energy and transportation are up double digits, and producer prices are up double digits, other consumer prices will follow.
The faster your cash loses value, the stronger your incentive to trade it for something else. That something else can be financial assets.
> It's not got loads to do with large-scale, institutional investments
For investors, particularly retail investors, the consumer price index is most relevant. But for whatever it's worth, producer prices are up over 16% in April (7% excluding "foods, energy, and trade services," which jumped over 50% annualized) [1].
To be clear, I'm floating a hypothesis here. I have seen no evidence linking inflation to demand for these companies' shares. (If anything, it should be the inverse.)
That depends. Inflation is a measure of the cost of living in terms of currency. It can be high either if goods and services required for living become scarce, or if currency supply increases. Currency supply increasing does affect asset prices.
Yes, but they're not directly correlated. Of course events can affect them both! Going to war would both increase the cost of living and (some) asset prices would go way up. But that doesn't mean they should be measured together like that.
Right now SPY not be such a great idea with SpaceX launch upcoming since it will be included into it immediately. Retail investors will be bearing that particular flop's cost.
I was old enough to remember the 08 crash. Then the market starting recovering in 2011/2012 and the sentiment was that the system would crash again soon like 08. Turns out, it was an amazing time to invest.
Post 08 crash, all sorts of conspiracy websites like Zero Hedge were popular saying how the world economy would keep crashing.
The only reason this happened was due to taxpayers bailing out financial institutions. This only exacerbated an insane amount of moral hazard already present in the market following previous bailouts.
Unfortunately, the US Government continued to run themselves into the ground spending-wise and may have a difficult time with another bailout, unless everyone pretty much agrees that we cannot have a USG failure, so they all pretend like nothing happened.
Eventually the merry-go-round stops, I'm just not sure what the catalyst will be, and it might be 100 years from now.
I am old enough to have had multiple career changes since starting on a major firm’s rates floor in 2008. These IPOs are tiny compared to the overall stock market and the stock market is absolutely tiny compared to debt markets. People consistently underestimate the size of the world economy or even their local economy. The world may look small from an orion capsule near the moon but almost every aspect of human society is bigger than most people can reason about. It is possible these IPOs have an outsized impact on sentiment for weird reasons. But it won’t be an actual outsized impact on capital markets.
Edit: I should add the AI bubble can absolutely burst but there is no reason to believe these IPOs are the end of the ride. If I knew I would be…
The thing is that the IPOs necessitate a full release of their actual costs for inference and training. This by itself should be enough to pop the bubble, if the occasional bits of it we get are anything to go by.
There is a reason anthropic is still hiding those details:
> key details typically included in that form about a company’s operations — like potential risks to its business, executive compensation, and other financials — won’t become public until later on in the process
The birthrate across all western nations has been below replacement levels for quite a while. All nations will see population shrink in the coming decades (aside from labor importation)
no one is going to get wealthy buying SPY/VOO. you might get rich, but not wealthy. things have changed now in a sense that handful of companies are large percentage of the stock market to the point where one has to question why invest in “s&p 500” vs “s&p 25-ish”
while going with the tried&true makes some sense, I think we have to open our eyes to a different reality of our stock market… and this market concentration into few companies is going to get a lot worse…
> things have changed now in a sense that handful of companies are large percentage of the stock market to the point where one has to question why invest in “s&p 500” vs “s&p 25-ish”
A small number of companies have always driven most stock-market gains. Betting on size isn't fundamentally a bad bet. But it is a bet against value and the historical tendency for small companies to be higher risk and higher reward.
you may be technically correct but today’s concentration in say top 10-15 companies is historic and by significant margin. I have been self-employed for a long time and somewhat “forced” into being “an investor” and starting in 2021-2022-ish I took my money out of all the “funds” … while I do not disagree that it is “a bet” - it is a calculated bet. things are different now even if historically you are right, no question
What do you mean by this? It seems pretty likely to be wealthy by investing in these indices. Certainly a “normal” worker who started investing $10k/year in SPY when it started in 1993 has enough wealth to allow them to retire now.
I very much disagree that it's coming. I think we need to completely reset our expectations of how the market works. There's been nearly an entire generation working in this "new" bull market, where things like EPS mean absolutely nothing and speculation no longer requires actual returns.
> the crash (that we all know is coming) will do that. Until then, history teaches that we'll just keep going up and up
Stock prices don't have to crash. They can just stagnate while profits catch up and multiples compress.
Debt binges, on the other hand, tend to go bust with a bang. But after the recent private-credit scare, the AI build-out has been predominantly financed with stock. (I think.)
Hasn't there been a _lot_ of debt to buy up Nvidia GPUs? I follow this stuff somewhat closely and it feels intentionally confusing, so I've likely lost track.
> Hasn't there been a _lot_ of debt to buy up Nvidia GPUs?
I believe that's been concentrated at the hyperscaler layer, and subsided when the aforementioned private-credit scare reared its head. (I haven't heard a big datacenter debt deal announced in a while. Though of course that doesn't mean they aren't being done.)
> Equity bubbles don't have to crash. Prices can just stagnate while profits catch up and multiples compress.
Is there is historical evidence for that? As someone who used to follow Jeremy Grantham a lot (he considered himself a "bubble historian"), IIRC every bubble he studied always mean reverted, and it usually (maybe always, can't remember) overshot on the downside during the correction.
> IIRC every bubble he studied always mean reverted
This really depends on how we're defining these things. Let's call a stock-market bubble a period of elevated multiples. That can mean revert by prices decreasing while earnings stay constant or by prices staying constant and earnings rising. (Alternatively, both earnings and multiples can rise and fall.)
Yes, for equity prices in particular he talks about P/E ratios (among some other metrics like corporate profit margins), and so you're right, it would be possible for this to mean revert by prices holding stagnant and earnings catching up. However, as far as I can remember (primarily because a big emphasis of his was how unchecked bubbles can cause a lot of damage on the downside) all the historical bubbles he studied (something like 50) always crashed with a big price drop. Not 100% sure though, which is why I was curious if you had any contrary examples.
Corporations across the board are experiencing record profitability. That's the reason behind the high valuations.
This isn't true of AI companies...yet. But these are companies entering the market with pre-IPO userbase (including lots of B2B) numbers that Meta and YouTube would have dreamed of before their acquisition/IPO.
I think this whole situation is very sleazy and corrupt, but ultimately my prediction is that nothing serious will come of it. Even the exposure of index and passive investing is overstated.
My first ever programming interview was like a group interview. There were three or four programmers asking me questions, one at a time.
The only one I remember was to check if two strings were equal (in C). I wrote (maybe buggy) code to iterate both pointers, comparing while looking for the null terminator.
The interviewer stopped me and said, “You should compare their lengths first. If they are different, you can exit early.”
I was pretty young and didn’t know much, but I explained, “But you have to look for the terminator to find length so it’ll take twice as long.”
He snapped, “There are optimized functions for that!”
I assumed he was right. Needless to say, I didn’t get the job.
Maaaany years later, I realized the std library was probably open source. So I checked (one). It was nice to be vindicated :D
C is pretty bizarre but I expect someone writing it professionally to know that even passing void compareStrings(char str1[], char str2[]) is equivalent to compareStrings(char * str1, char * str2) so no way to get the length of it with sizeof(str1) and strlen walks the string until it finds the null terminator.
No, it's not bizarre, it's standardized on what is and isn't UB based on history, and usually for performance.
NUL-terminated strings don't know their lengths and so, without an "n" variant function and running strlen() ahead-of-time, must iterate the entire thing. Pascal format strings (supported up to 255 byte lengths in the classic form) could find length as an O(1) operation because there was no terminator necessarily.
"Group interviews" produce horrible, unfair dynamics.
Snapping with know-it-all arrogance is toxic and doesn't help anyone. You were correct because strcmp has to iterate both strings, just like strlen would.. and it's totally pointless.
#include <string.h>
/* C89 and handles NULL and overlapping strings */
int strequal(const char *a, const char *b) {
return (a == b || (a != NULL && b != NULL && strcmp(a, b) == 0));
}
I think you dodged several bullets by not getting hired there because they sounded insufferable.
> Pmf is this weirdly defined thing where "if you're not sure you have it then you don't".
I'm not sure if this runs counter to your point or not, but: I don't see any future where LLMs aren't a core part of Software Engineering. The horse is out of the barn. There is no going back.
Yeah but the product is not “LLM” it’s “proprietary frontier model LLM paid by the token”.
And I don’t even necessarily disagree with OP! It’s more like the competition is shifting so quickly that your competitors could undercut your PMF in a blink of an eye.
History bears out that cheap and satisficing soundly beats expensive and optimal every time. Until we have smarter and more prescient decision makers in leadership, the bottleneck on output will be the quality of decision making not the quality of code. Trying more things faster and cheaper will win.
The frontier models know Norwegian just fine. They can also adapt to Norwegian dialects, and even ape old Norwegian fairly well.
E.g. I had Claude describe the novel "De knyttede næver" from 1911 in Norwegian orthography ca. 1911, as it's a novel I've read, and it does a good job.
What it lacks is an understanding of Norwegian literature, culture and history. It had to look up "De knyttede næver", which was one of the best-selling Norwegian novels around the time it was published before I'd get anything out of it (ChatGPT does better; in thinking mode in particular it gives a detailed summary).
While not exactly well known today, the author was a prominent newspaper journalist for decades, and the novel series is well enough known that e.g. there's a Norwegian singer that took his stage name after the protagonist, and it was covered in Norwegian papers and books for decades (partly because of controversy over the authors political views and how they coloured his novels), so it does feel like a reasonable test that reveals a quite significant knowledge gap.
I do agree with you that it'd be better if the data set from the national library was made more accessible, though it seems a major addition here is that they have a deal to train on copyrighted data locked away in their archives that they have limitations on the use of.
But even just making the out of copyright data in their collections would be a great start.
You'd think so. It seems like there are a lot of odd gaps like that.
I also have a favourite English language PhD thesis I ask every new model about that they still struggle to find even though there's a Wikipedia article about it that links a blog post I wrote about it.
Anyone who thinks they've exhausted even publicly crawlable resources should ask them about some obscure stuff.
you might be surprised if you take this approach.. give key words and phrases in small amounts, each sentence of a prompt building on a previous sentence. Take a an example that is not very hard, like Lewis Carrol Alice in Wonderland original text. Although a quick question might get things sort of wrong, or miss details, if you guide the LLM to a certain part of the story, then a certain set of characters in that part of the story, then a certain statement or dramatic moment with those characters in that part of the story, you might get very specific detail that is close to line-by-line accurate. On the other hand, if you ask a quick, ordinary question about the same part of the story without supplying context and character names, you get something equally vague. YMMV
For the PhD thesis in question, I've actually tested a lot of requests about different parts of it, and both Claude and ChatGPT still draws a total blank if you don't let them do searches.
Why should they share all this data with the greedy american corporations that are stealing everyones data for their own profit? Much better to keep the legal agreement with the national institutions and possibly develop something actual useful to their own country.
You are contradicting yourself. If you're hoarding the data for yourself you're not going to develop something useful. Sharing the data means that it will be integrated into the big LLMs, which will be useful "for their own country".
Obviously, it's fine to be wary of any development in your area. But it seems like there is a certain amount of irrational(?) fear of datacenters. And I really don't understand it.
I saw a poll recently that people would rather live near a nuclear power plan than a datacenter. That's... their choice, of course, but doesn't seem logical to me.
I have heard several "concern stories" about them on NPR recently. Maybe there is a political component to it. But I do worry there is some kind of manipulation being done.
> When low-frequency sound becomes strong enough to be heard or otherwise felt, it can cause annoyance, discomfort, and sleep disruption like any other normal noise pollution.
So which is it? Sure, I don’t really believe that there is magical super special harmful noise from a datacenter, but are these monster datacenters emitting disruptive amounts of low frequency sound or are they not?
Ad homniems aside, is the accusation even accurate? So far as I can tell he doesn't obviously have "a massive financial incentive to do so", like he's a VC investor in anthropic or whatever. He does seem to be bullish on AI in general, but I'm not sure why that'd be a disqualification for someone on the pro-ai camp any more than someone who's interested in retaining their property values or whatever would be a disqualifier for the anti-ai camp.
Ad hominem would be if shimman had said something like "don't post rebuttals from people who are stupid meanyheads". Identifying a characteristic of the posters that affects their incentives is a perfectly legitimate reason to discredit their posts, or at least call their impartiality into question.
>Ad hominem (Latin for 'to the person'), short for argumentum ad hominem ('an argument to the person'), refers to when a speaker attacks the character, motive, or some other attribute of the person making an argument rather than the substance of the argument itself.
I think it's as simple as people generally believe that electricity is good for the world, and AI is bad for it. In the former case you're kindof taking-one-for-the-team. Even places that have nasty nasty stuff like tailings ponds generally have a kind of civic pride that coal mining or whatever was a necessity and the sacrifice of their local environment made a lot of other people's comfy lives possible. Data-centers are just not going to inspire that sentiment lol.
For some strange reason people aren't all that keen on building something that'll increase their utility bills, pollute everything, and threatens to take their job.
Here in Michigan, people are all for auto factories, but the polluting, energy intensive, and job taking data centers are a big no no. They use electricity, they look ugly, and they use water. Can't have that.
> Here in Michigan, people are all for auto factories, but... data centers are a big no no. They use electricity, they look ugly, and they use water. Can't have that.
Because they know economics better than their politicians and academia.
Data centers saddle the public with their power and water capital expenses, for new generation and transmission which are used solely for the benefit of data centers. And get this, in many cases the data bros get sweet tax-free deals for many years.
All of a sudden, the entire economics establishment loves communism for the rich, where the rich get exclusive use of public utilities built and paid for by the public.
The media, academia and politicians silence is deafening, which is why people have to raise their voices if they want to be heard.
Thanks to information campaigns people who live near nuclear facilities tend to have an above average, positive, view of the safety and threat.
A large part of my extended family lives near a large facility. Wind turbines launching ice at the nearby roads is a larger (yet trivial), safety concern.
I don’t know the background to this project, but a nuclear project would likely be very transparent - with public studies on the impacts and meetings for the public to make their views known. It’s far quicker to build a datacenter than to increase local grid and water capacity later.
> The Stratos Project moved forward with far too many unanswered questions around water, power, cost, and transparency.
In a town near me a paper plant recently closed (to much anger), then people protested the potential use of the land for a data center, citing concerns about noise, water use, power use, and traffic.
I have childhood memories of visiting my grandparents, who lived near a town with a paper mill. Going to town, going to a restaurant etc, meant being inundated with the horrible smell produced by that mill, within a radius of miles. It was a fact of life there. Genuinely hard to imagine a data center producing worse waste products than that.
Datacenters are financially a net negative for whichever municipality they end up in. They're operated mostly remotely with little staff and they have no tangible production, meaning any wealth they generate ends up vast distances away. Meanwhile the municipality ends up with increased costs because of the inefficiencies of bruteforcing computation, and because of the subsidies and tax breaks that the companies not only expect but demand for construction, there's no revenue being generated even for the local government.
Because in reality they don't actually bring in any revenue for the first few years thanks to all the subsidies and tax breaks they demand upfront before construction. At best it would be state and federal taxes used for operations that any operating business brings like federal payroll, but there wouldn't be any property taxes for at least five to ten years and the federal corporate income tax would likely be from the state the company is based in rather than the state the datacenter is based in. The municipality, be that the county or city the datacenter's in, gets screwed.
Meanwhile just to run a trucking depot you'd have the heavy vehicle tax, international fuel agreement tax, registration tax, sales taxes for the trucks and trailers, property taxes, and whatever incidental taxes required by the state you're operating in. The property tax, IFAT, and local payroll taxes meanwhile all go to the municipality and don't skip straight up to the state or national level. This is with no expectation of any of this being waived or delayed because the trucking industry doesn't have the surface visible financial performance of the industries municipalities are more lenient towards.
> Over the past 20 years, the data center industry in Loudoun County has grown significantly. As a result, the amount of revenue that the county receives from personal property tax revenue on the computer equipment located inside data centers has also grown significantly. Currently, data centers occupy approximately 4 percent of commercial parcels of the land in the county; however, they yield 38% of general fund revenue collected by the county
38% of fund revenue for this county comes from data centres.
It’s pretty simple. People think that AI will take their jobs and maybe murder them, probably because the people developing AI have said it’s going to take their jobs and maybe murder them.
Opposing data centers is the biggest lever most people have to impede AI development.
They're big, use up a lot of power, destroy a large batch of land, produce noise and locals get basically nothing out of that (it's not like they provide a lot of jobs or anything). The power bills also go up.
As someone who lives in Northern Virginia, it makes me furious to receive my electricity bill and see that even though I used less electricity than the same month last year, I am paying significantly more. And this happens every year.
Do you think Virginia is adding solar, battery, and wind proportional to that additional power draw? Nope! It's natural gas and coal power imported from PA and WV. It would be one thing if I was paying more to build out renewable energy for environmental purposes and to set up a reliable and clean grid for the future. But no, I'm just subsidizing these huge companies and hurting the environment to boot.
This echoes some of my biggest gripes about data centers:
We should be mandating green power, to a great extent, be built to support these facilities.
We (US states) should not be competing, in a race to the bottom, to be the state to give the biggest tax breaks and pass the cost to the citizens.
We should not be ignoring the citizens who will have their health and livelihoods affected.
AI data centers, for better or worse, are very necessary for many reasons. They could be built responsibly, or at least less hazardously, but the care isn't being put into that aspect of their construction.
So true. A legislator in Virginia finally pitched ending the tax credit that brought all the data centers to VA in the first place, and I hope it passes (I know it won’t). But seeing the upcoming rate increases already on the books and the number of data centers they are planning to build is pushing me to consider solar again. The payback time is getting shorter and shorter :)
Aside from "moral outrage" style concerns ("AI is bad for the environment", power consumption, water consumption, or "datacenters benefit rich people, rich people bad, so datacenters bad"), I've heard of specific bad examples how datacenters (allegedly) negatively impacted the surrounding population:
- Noise (from fans to generators to possible infrasound concerns)
- Air pollution (from data centers semi-permanently running on generators)
- Electricity prices (although I don't understand how this is supposed to work)
- Water consumption affecting the population (water restrictions, price increases, water table dropping)
Many of these are one-sided stories told from the perspective of the residents that I didn't try to verify, but I suspect some of these concerns are legit.
The company building the datacenter has a lot of incentives to cut corners and/or cause some of these impacts, externalizing its costs (e.g. by saving money at the expense of noise emissions, running the DC on unpermitted gas turbines to be able to build a DC where there isn't enough grid, negotiating clever deals that benefit the company but screw over the utility forcing it to raise prices for others, using groundwater for evaporative cooling to make cooling cheaper, etc.)
The company building the datacenter also likely has a lot more experience while the people of the town and the town itself are doing this once, so there is an inbalance in experience that makes it easy for the company to get away with some of these.
There is very little benefit that the people of the area can expect from a data center - as I understand it, there are very few jobs in one past the construction phase, even the construction jobs are often filled with experienced travelling workers, and given the negotiation imbalance, a town seems likely to get screwed on any contributions that the data center promises.
Maybe the solution would be some kind of framework/organization that guarantees (ideally with binding, well tested contracts) that the datacenter won't be a nuisance, builds a reputation for being reliable, and in exchange, companies that work under that framework can expect quick approvals and less pushback.
Until that exists, or companies start offering guarantees up front (e.g. guaranteeing a certain power price or noise level), I'm not surprised that people push back (especially if the company building the data center has screwed up in the past).
All of these concerns are valid. But none of them are unique to datacenters.
A golf course uses a lot of water. A factory can use a lot of power -- and generate pollution. A chemical factory could have all kinds of externalities (if not properly managed.) Heck, switching to electric heat (over gas) or electric cars over ICE for an area will also drive up power usage.
But we don't freak out when someone builds a golf course or a factory or switch to electric.
We have rules about all those things. Sound is one: you need to be within reasonable limits. Electricity usage is another: power operators always need to manage their load and expand generation (that's why we keep adding solar and wind everywhere.) Air pollution is similarly managed.
I can understand if people are concerned about "infrasound" -- why not pass a law that regulates it -- like other noise limits?
Datacenters may have specific potential issues. But none of them are unique to datacenters. And we've been managing these issues for hundreds of years.
> I saw a poll recently that people would rather live near a nuclear power plan than a datacenter. That's... their choice, of course, but doesn't seem logical to me.
The resource consumption is huge and it provides relatively little to the surrounding community compared to its intake. For most residents who live near one it’s a net loss. Qualify of life decreases and utility bills go up so that a Silicon Valley exec can get a nice bonus for closing the deal.
A nuclear plant creates energy and a decent amount of jobs, while a data center’s value is dubious to the average human and the data center barely brings in any jobs.
> I saw a poll recently that people would rather live near a nuclear power plan than a datacenter. That's... their choice, of course, but doesn't seem logical to me.
Yay people have finally become rational about nuclear power safety !!!
It's not clear from the article whether this is just datacenters -- or if that is just the convenient boogeyman.
The grid operator for the northeast, according to my Governor, has been well-behind in building out infrastructure. Of course new datacenters cause more load. But so do new houses (we're building as many as we can) and electric cars, etc.
It is not entirely wrong. Annual electricity consumption per capita in Maryland has been falling since 2005[0]. Unless a bunch of aluminum smelting plants have come online in the past 18 months, data centers do seem the appropriate root cause leading to a surge in demand.
There’s another chart on that page that shows regional demand for electricity. That seems to have flattened or dropped. Not sure how to explain that, in the context of data center demand.
Maybe I’m reading something wrong. Or maybe there is an anticipated increase in demand?
If you click on the edit icon, you can see more data. It show 6 months worth. I thought it was longer. My mistake. But I guess I would still expect to see something over the past 6 months given the hysteria?
Because we have offshored most of our heavy industry. For national security reasons we now have to re-industrialize. Plus a lot of the infrastructure for transportation and heating is gradually shifting off of fossil fuels. So electricity usage is going to go steadily upward for the next couple decades. This would be necessary even without AI data centers.
I think we reached peak "electricity" for the consumer decades ago. You know...when our TVs were a giant electron gun pointed as a phosphor screen whilst the room was lit with a filament of 3000 degree tungsten.
...and Americans were getting like 200HP out of a whole ass 7.whatever L v8 ahaha.
Is it fair to Dutch citizens that they're going to be left behind in what is effectively the next industrial revolution? The Netherlands does pretty well in other economic sectors including fossil fuels and agriculture but those aren't going to enable much future growth.
There is so much FUD going around SPECIFICALLY about data centers lately, that i’m dubious of anything i hear. it’s such a weird cultural phenomenon. Chronically online teenagers on Instagram making increasingly incorrect and absurd-sounding claims about water / energy usage. Comparatively barely anyone knew what a data center was 1-2 years ago.
The pitch for AI from its proponents is the destruction of all white collar and creative jobs, while simultaneously making all the content they consume worse and Internet socialization with human strangers replaced by doppelgangers? And you wonder why people are upset?
I believe the proponents think it will make all the content better, not worse, but the current state of affairs is indeed definitely worse.
But I essentially agree with you, the destruction of all white collar and creative jobs is something the proponents hand-wave too easily with vague mention of UBI as a thought-terminating cliché rather than a real policy proposal.
(Myself, I'm in a strange position with doppelgängers, because I simultaneously want real human connection and keep getting disappointed with many of the real humans).
> (Myself, I'm in a strange position with doppelgängers, because I simultaneously want real human connection and keep getting disappointed with many of the real humans).
I don't think you're in a strange position at all. You are just being honest about what is the undercurrent of much of modern technological development. AI is simple the apex of a trend to replace human interaction with a simulacrum, which without your awareness, reduces your own tolerance for the frustrations of real human interaction, thus making you more dependent on the simulacrum.
> I believe the proponents think it will make all the content better, not worse
I believe the proponents think it will make content more profitable, especially in the sense of concentrating the wealth that can be extracted from content creation. Related, there’s a desperate attempt to redefine code as “content”, at least implicitly, so that it can fall under that umbrella as well.
There's pretty noticable differences in ability to self-regulate between people who had unfiltered access to computers and smartphones when growing up and not, beyond simply generational shifts.
I think the parent's assumption is that 16 is old enough for those problems to be largely avoided even if you gave them unfiltered access, which is what makes it not just kicking the can down the road.
I know a six-year-old with an iPhone. Any efforts to not do that is going to help.
I remember early internet and how it was on a teenager. Can’t imagine just signing up your kids for what it is now just because I won’t be able to keep them from it forever.
The draft term is 'computational load'[0], which is the focus of the L3 essential action in that press release.
NERC's concern appears to be computational load cutting. In 2024 for instance, 60 data centers in NoVa cut 1500Mw of load in a matter of seconds[1]. Dominion reported 70GW of large load deliver points, with 40GW in new requests that are 'vast majority' data centers.
40GW is approximately enough to power about 30 million homes (at 1000 kwh/month).
My guess is that it’s a Chinese PR campaign to sew bureaucratic barriers to data centers. No evidence whatsoever, but it seems to me that every social media driven trend that benefits some corporation is a pushed idea.
Somehow I don't see arguments comparing powering AI with producing food to go over well with them.
Many opponents to AI do not view the tech as having a net benefit. Comparing it to food production would serve to make you look more the fool to them despite their claims about water consumption frequently being wacky.
I am finding that 42% of California water is agriculture and 8% of that is for almonds. 3% of state water usage on a luxury nut might not be moving the needle that much, but it does feel wasteful for a state that is perpetually hovering on drought conditions.
It's even more wasteful when you realize that the USG is trying to force governments around the world to buy Californian almonds, even in places where almonds are traditionally grown and harvested.
And then there's the matter of Saudis coaxing farmers to grow alfalfa which is then exported to Saudi Arabia to feed cows there (!).
Precisely my point. The US grows an unsustainable amount and then pushes it to places it was grown in traditionally like the Mediterranean, Western Asia and the Middle East.
The USA produces the lion's share thanks to USG subsidies, then the USG comes over to the old world countries and the commerce reps start faulting the countries for a.) blocking American agricultural imports or b.) subsidizing their own local agriculture.
Since the American almonds are also produced at a much larger scale, they can also underprice the locally produced almonds ridiculously, which further disincentivizes local farmers from growing almonds and forces them to grow something else.
Right; anti-data-center sentiment is really a way of attacking AI as a technology; arguments about the water or power use of data centers are just an excuse.
Well, to be fair, the public has been going through a multi-year forced beta-test of AI all while CEOs keep going on national television and posting on social media how there will be mass unemployment because of AI. To say nothing of all the companies that have (and soon will) close up shop due to increased prices from the global memory shortage unrelated to the proclaimed job-replacing benefits of AI.
And let's not forget many of the remaining independent websites on the Internet closing up shop due to being unable to afford the substantial increase in hosting costs resulting from aggressive scrapers getting data to keep training AI on.
Or the massive improvement in bots and click-fraud due to AI, pushing an increasing number of companies to embrace heinous practices such as mandatory facial recognition for users to be allowed to engage in socialization.
Or the increased electricity prices already realized around much of the country due to AI both so much of the existing grid's supply and requiring expensive upgrades to the infrastructure - the latter of which is frequently paid for by taxpayers.
All for the wondrous promises of unproven future capabilities.
The real mystery to me is how it's a mystery to so many people why there's a large and enduring anti-AI sentiment.
Yep. 10000 gallons per query, spews out toxic water, contaminates the water supply, uses more power than an entire state, and on and on. I'm convinced it's a psyop to prevent the US's progress in tech. It's so over the top crazy and obviously false, but everyone I know is falling for it.
They might be wrong on this one, but check the history on cancer alley, or the legalized massive PFAS dumping in rivers the world over, that has now polluted the earth so thoroughly you cannot escape going over the maximum recommended body serum. Same for microplastics.
It is absolutely justified to be extremely suspicious of big corporate. They've earned it.
Understandable and justified are very different. I worry that unprincipled skepticism of big corporate makes it harder to stop the bad stuff; if every large project becomes a battle of corporate power vs. slopulist criticism, how do you sort through that to focus on the truly bad ones?
Legacy data centers were not representing a multiple of existing electricity demand. It is only recently that DCs are being built hoping to consume hundreds of MWs, up to GWs. Elsewhere in the thread is the Utah DC which will consume more electricity than the entire state.
Electricity demand growth in the country has been flat for the past twenty years. DCs are causing real strains to the grid which has not had to accommodate rapid growth.
Seeing a datacenter proposal in the news come in at 1.3GW was very sobering for me. I spent a lot of time in grad school on the campus of a large nuclear plant, and it turns out one nuclear core is good for about 0.9GW of electricity (or 2.9GW of heat).
A single site consuming more than the entire electrical output of a nuclear core, considering the sheer size and scale of that reactor and its supporting infrastructure and workforce, is just boggling to my mind. It's literally billions of dollars just to feed that one site, if they're being accurate in their proposal.
It seems pretty easy to undermine the trust in the other side if you just pretend to be on it and inflate their numbers. See: Florida will be underwater by date X [in the past]
‘More power than an entire state’. Yep - take the Stratos data center project in Utah, the first phase of which is expected to consume 3GW and at full capacity is expected to be 9GW. By comparison, the entire state of Utah currently uses about 4GW.
For a less rural example, I am finding different numbers, but all of New York City is estimated at somewhere between 5-10GW. So, some 9 million people vs one data center.
That's a take. But let's stop and ask ourselves -- Cui bono?
Another take is that the same companies that are pushing for datacentres are often the same companies that control social media and traditional media outlets and are using this control to foster datacentres onto thee average person who is either wildly unenthusiastic about or at best ambivalent about.
It's all pretty moot anyways.
Big tech oligarchs have gotten pretty much everything they want over the years, it's not like the average person in bum-fuck nowhere is really going to be able to stop them from destroying their watersheds, poisoning their air and jacking up electrical prices.
I wouldn't get too upset about opposition to datacentres if I were you.
Money is King and the King has spoken.
There will be datacentres where ever the tech oligarchs want there isn't anything anyone can do about it.
Honestly, I hate "free tiers" for SAAS products like this.
For about 5 minutes, Google had a pay-as-you-go service on GCP for git. I used to use that, because I wanted to own my stuff. But (I guess) because everyone used "free" github -- and like a lot of other Google services -- they sunsetted it.
So now I'm using github for free. But would rather pay for my storage an usage with a (big) cloud provider.
A terminal with AI focused on doing terminal-ish stuff is actually kind of useful.
I just never did enough of it to keep going.
If they expanded this to be highly optimized for devops aka really well attuned to AWS CLI all the various linux commands, bash scripting and just had all of that baked right in - and - was super fact and didn't have to think to much - I can see that.
The reason being, your doing 'specific tasks at a meta level' - not designing complex things, or doing research.
More like Claude Code but not for code, for DevOps and or that kind of things.
I think 'Meta Prompting' should be a thing for many disciplines.
That said, the 'bitter pill' lesson is that the Tier 1 models just really get good at everything and often supersede custom solutions - which was the case for myself and Warp, I just 'did stuff in Claude' and it was good enough.
Claude Code is very capable of making a terminal emulator with exactly (and only) the features you want. I did that for myself and it's now my daily driver. Has a few goodies I care about but nothing much else, and I have no intention of adding features for other people: https://github.com/cartermp/term
Maybe getting more of these big private companies public will bring valuations down a bit.
(Just my impression. No math or financial studies behind it :)
reply