I mean... You can buy a stick of RAM with Samsung chips on it and use it in your own machine to run a local model vs. paying for an Anthropic subscription. I assume that's what they're referring to, not owning Samsung stock.
I've commented about this before, I think many LLM based apps nowadays are at risk of being replaced by a product straight from the labs once they prove to be successful.
We've seen this pattern with Apple making their own version of an app that was previously popular on the app store.
The labs are in a perfect position to do this - they have a bunch of data on what's being used and they have direct access to their own models/compute. If an external service is popular, it's relatively trivial for them to estimate how much additional profit they're leaving on the table.
If the usecase isn't far from their core business (and things like this absolutely aren't), with their size, why wouldn't they eat other people's lunches?
They do use water, and the numbers look very large which makes for big scary headlines. However people who actually know how water is used in a city are used to numbers much larger than that.
it's the demand delta, not the total demand, that's concerning. Also not all use cases are created equal: using water to keep people alive is legitimate and using water to build a machine that can automatically almost count the number of vowels in a sentence is not legitimate, especially when that usage drives up demand and makes it harder to use water to keep people alive.
Well "you" don't "get to", but this is precisely why we made the concept "states" and use "state power" to counteract other natural modalities of power.
Which is being demonstrated by the use of state power to kill data center development.
The difference is they already exist while the datacenters don't. And its a lot easier, in a democracy, to stop a thing from being built in the first place than to tear it down after the fact, what with pesky laws about private property and so forth.
Or are you really saying that if we can't solve all water usage problems then we shouldn't solve or prevent any? Is that what you're suggesting?
It is deeply funny to me that my asking a rhetorical question and using a little italics triggered you while "So, you can scream fake news all you want, but..." earned not a single complaint.
I would expect that a golf course construction is relatively rare event. And there is definitely no explosion of new golf courses. Total DC power consumption growth is practically exponential for the last 5 years.
>I would expect that a golf course construction is relatively rare event
That's just status quo bias, given that they're an ongoing source of water consumption. "oh it doesn't matter that [thing] is guzzling water, because it's already there!".
Exactly right. It's a lot easier to prevent an asshat venture than it is to dismantle it. Now that the public is aware, I wouldn't be surprised if new golf courses start getting the same treatment.
They haven’t been. I’m aware of new golf courses being planned in my area and I hear nothing but enthusiasm. Literally no one brings up water consumption as a concern.
This whole thing is a sublimation of vaguer anxieties about AI (which are perfectly rational!) into areas where people feel like they can make more concrete objections that are likelier to succeed. The water thing is completely fake, whether people think so or not.
Not entirely. Lack of regulations mean recycled water facilities can end up being consumption facilities with pollution risk. Also Meta plans to restore as much water as they use by 2030. Which, seems like a while to break even. That's primarily because ai data centers require a lot more cooling water and power than traditional ones. So, you can scream fake news all you want, but it is a concern and it is a good way to oppose ai in general. And power consumption, noise pollution, and pollution are also serious issues. "Whatabout this worse thing" is a pretty terrible argument. You can whatabout anything, but it doesn't make that thing any less bad. aka two wrongs don't make a right.
It’s not just whataboutism, it’s pointing out that the people up in arms about data center water usage don’t even know about the long list of things that use way more water than data centers, and when told, they don’t care.
I think it’s pretty reasonable to conclude that their beef is not really with water usage as much as a general hatred of AI, big tech, etc, and water usage is an easy thing to focus on.
golf courses don't contaminate the water supply. Amazon was just forced to settled a lawsuit for 20 million because they took a community's water supply and made it undrinkable. Now they'll be paying to improve public water treatment infrastructure and dig deeper wells looking for cleaner water. When's the last time a golf course had to do that?
Yeah, it’s one thing when SV runs the Uber script (ignoring the law to build big enough to buy politicians) to hurt cab drivers. It’s very different when it’s your neighborhood.
Why and how would they contaminate the water supply though? Sounds like a municipality problem, a water consumer shouldn't be feeding back into the system, or is it being caused by poorly executed construction practices? Maybe when they first pipe in or something is being left open during construction, this seems like it shouldn't happen if things are being done properly, water supply systems are not a new technology so who is dropping the ball?
Yeah, all these stupid little municipalities have to do is shit they've literally never had to do before while getting sued out of their socks by literal trillion dollar companies.
That's all it takes for them to do their part in generating gargantuan amounts of value for people who live 1,000 miles away and their uberwealthy hardly-human CEOs and investors.
it's social media, and a human who strongly signals they know something is usually much more reliable than a search engine for knowing the real upshot, 9 times out of 10 the person who already has a strong opinion has keyed in on the detail that even if I found a search result might not be so easy to decipher, or you could have some professional experience, anyways it was worth a shot, thanks for sharing
In the case of one amazon data center the town already had nitrates in their water due to agricultural runoff, but at levels that weren't a problem. Amazon took in hundreds of millions of gallons of that water for cooling, which evaporated away some of the water, but left the nitrates. Then amazon dumped the concentrated nitrate 'waste water' at such high volumes that it was pushed so deep into ground that it reached the aquifer contaminating the source for some 45k people. Soon people's taps were delivering water contaminated at 10x the legal limit. People got sick. A bunch of parents lost their babies.
Amazon was forced to settle for 20 million which will pay for digging to find uncontaminated water sources deeper in the ground. They'll also pay for water treatments to help remove the excess nitrates.
Data centers don't have to bring in new toxins and pollutants. They can also make existing problems much much worse.
This doesn't even make any sense. "deeper in the ground" is mostly wishful thinking. Why wouldn't the contamination spread? How can anyone be sure there is even potable water deeper?
With regular rainfall/usage the nitrates don't get far enough down to reach the aquifer. The nitrates get filtered naturally and remain in upper layers of the soil. Amazon concentrated the contamination and drove it far enough down that it reached and contaminated the aquifer, but below that the father down you dig and pull up water the less contaminated it will be. Part of the settlement is a fund for digging new private wells deep enough that the water contamination is under legal limits. I guess if there's no clean water below the contamination they'll have to stop using wells entirely and depend on treated water from the contaminated source.
The construction itself absolutely alters local watersheds:
1. Replacing forests and fields with concrete buildings prevents rainwater from filtering into the soil, which stops local aquifers from recharging.
2. During storms the rain runs off impervious surfaces and can cause downstream flooding.
3. Heavy land grading washes loose sediment into nearby waterways, destroying aquatic habitats and turning private drinking wells muddy. (That especially gets local homeowners angry and gets lots of video views).
And then you have the ongoing operation of open cooling systems:
4. Evaporative cooling systems discharge concentrated mineral brine that spikes the salinity of freshwater ecosystems.
5. Cooling water is treated with toxic biocides and anti-scalants that can contaminate local water basins. It can also contain leached copper, zinc, and PFAS.
6. Discharging warm wastewater depletes oxygen in streams and alters the overall ecology.
Of course all of these issues could be mitigated to a reasonable degree with proper care and effort:
>this seems like it shouldn't happen if things are being done properly
Do you want to bet your whole town on things being done properly right now?
All of this applies to all construction of all types. There are already plenty of environmental regulations around buildings and their construction. Maybe these aren't being enforced, maybe the laws need to be stricter. I'm not an expert but it sounds plausible. But if that's the case, we have to fix those problems otherwise you're just playing whack-a-mole with development. When you specifically focus on datacenters and stopping datacenters it gives away that environmentalism is just a fig leaf for your actual goals.
The first three apply to any construction project with a paved surface and there are ways to mitigate that, you can collect the runoff in a cistern/tank that filters back into the groundwater or collect it in a retention pond which does the same thing.
Items 4 and 5 apply to any facility with cooling towers, not limited to but including hospitals, apartment buildings, schools, prisons, cold storage warehouses, combined cycle power plants, large commercial buildings, paper plants, retirement homes, commercial bakeries, industrial facilities with heat rejection needs, etc.
The sixth one applies to virtually every power plant in existence.
how would this mess with a water supply though, or are we talking the private wells? with a private well though it's something you want to mitigate either way, filters, etc, granted I have only seen the handful of clips where someone is showing dirty water coming out of the tap and I am trying to trace the chain of responsibility for the water supply, thanks for the detailed response btw
My county, for example, only processes surface water. If someone altered our upstream collection or polluted the reservoir, it would be a very big deal.
For private wells: if you're a rural person who previously had perfectly fine water, and now someone comes along to mess with that, even if it "just" requires installing expensive new equipment, you're going to be pissed.
Not sure why this is getting downvoted, most of the new facilities (e.g. Stargate) are running closed loop systems. PUE goes up a bit if you have to use chillers, but ultimately it's faster/easier to build a large facility in a closed loop fashion than get water (permitting, treatment, etc.). NVIDIA pushing 45C inlet temps is to allow facilities to run dry coolers instead of chillers and make this even easier! Two phase cooling (required above ~3kW TDP) would eliminate water use in the secondary loop (if not the primary, eventually).
Anyone else worried about the premise of Anthropic selling specific agentic tools that also use their models underneath? My first thought is apps on the iPhone that got banned after Apple introduced an official option with the same functionality in a new iOS.
I'm not sure I would risk building an agentic app that competes with this on Anthropic models out of fear of that scenario.
Is this some kind of trick question or am I just misunderstanding your reply?
As far as I know they sell the patterns, not the finished knitted artifact. Aspiring knitters buy them to enjoy knitting. The physical representation of the pattern isn't worth anything (just a piece of paper), what you're buying is a license to use their intellectual property, so I'd say they're making money from copyright.
Full-stack web developer with 20 years of experience, currently working at a startup where I was employee #2.
My strongest area of expertise, if I had to point one out, is frontend. I've built complex multiscreen dashboards, mobile games, IPTV set-top box interfaces, you name it. Consistent and performant UIs are my jam.
I enjoy working with people, I've happily and successfully led teams, conducted dozens of technical interviews and managed cross-team initiatives. That said, I'm a deeply technical person at heart, so I'm not looking for a full-time management position.
Over the past year I've been working with LLMs extensively, building an AI asset management product at my current employer, which required diving into coding harness internals and shipping features supported by LLMs. Our development processes started relying on LLMs as well, so I'm well prepared to join a team that's already heavy on AI or help it make that transition.
My preferred environment is one where I get a seat at the table early so I can own a feature from conception to production. I'm used to low overhead, frequent exchanges with product people and daily deploys. If you're not there yet, I can help you get there.
While I agree with the general sentiment, the problem here isn't developers not being familiar with statistics, it's the simple fact all of this is profit driven most of the time.
I tried to purchase tickets for an event last week. I had to go through Ticketmaster as it was the only official way. They forced me to verify my account using a phone number, but whoever they were using for messages wasn't able to deliver a code to my number. I tried a few numbers from our household and they all failed.
Searching for this issue yielded a bunch of results, so it was definitely a known issue, but there wasn't anything I could do, really. To them, it's simple math. Another SMS provider that covers my (tiny EU) country might be more expensive. They might be avoiding scammers that used my mobile operator in the past. Whatever it is, it would probably cost them more than they lose in ticket sales.
Without some government entity to force them, they don't give a shit about me being able to see an event.
Infrastructure should not be (purely) profit driven. To improve profits for train operators, the simple option is to cut lines serving small and rural communities. The economics are much worse than serving large cities. Same for cell coverage and broadband internet. Most profitable is to just not cover a few percent of the population.
There is a point where technology becomes foundational for participating in society. And then it needs to be regulated to be available to everyone.
Meanwhile in Britain in the 1960s, this cost-cutting closure of local rail lines did happen: https://en.wikipedia.org/wiki/Beeching_cuts ... at a time when the trains and rail infrastructure had been publicly owned for about 15 years already. It doesn't dispel the incentive.
I’m not arguing a rail has never been closed. I’m arguing that being a small difficult market doesn’t exclude you from being served by marketed forces.
Did nobody ever operate rail to those cities again due to them being rural?
Rural cities? Come again? What was demolished remained demolished, yes. Unclear on your point.
Oh I see (thanks to that edit). I mean, I agree with you. This is just the additional amusing detail that government-run services are still subject to a sort of dulled and homogenous version of market forces, which can be worse for small local concerns because it's less responsive. Though, admittedly, a giant corporation can simulate government very well, and can be just as crap.
> A few of these routes have since reopened. Some short sections have been preserved as heritage railways, while others have been incorporated into the National Cycle Network or used for road schemes
> Some, such as the bulk of the Midland Metro network around Birmingham and Wolverhampton, have since been incorporated into light rail lines.
Furthermore, the transformation to other transportation forms suggests this event also coincides with changing technology.
I have never heard the name Beeching spoken with more venom than in Wales, I used to live in Mid Wales and now I live in Cardiff. If I wanted to visit where I used to live by train, I'd have to do a multi-hour detour of a sightseeing trip around the West Midlands, deep into England. The Beeching Axe literally cut Wales in half and the consequences are felt to this day, even though there wasn't much outright salting the Earth to make sure the terrible decision couldn't be reversed as there was in some cases, the Welsh government doesn't have the money to reinstate the Aberystwyth-Carmarthen line which would deal with a lot of these north-south issues.
Also it's not just Wales where Beeching carried out intense vandalism of public infrastructure, the South West was severely affected too. Basically anywhere that wasn't London-centric suffered, which is the British government to a T regardless of the party in power. The general assumption was that private cars would replace the local trains, which as someone who currently doesn't drive for medical reasons really makes my blood boil. While perhaps not in intent, in effect the Beeching Axe was a profound kick in the teeth for the disabled.
You cut off the OP's sentence of that being examples for "Rail companies" and then added your own examples. Please be better at comprehesion and editing comments
"In practice Amazon will deliver packages across the US." You know they use the Postal Service for last miles often? And the Postal Service is required by law to service far-flung places. So Amazon is only, in practice, delivering packages to those places due to USPS.
The airline industry gets huge subsidies in most countries to operate more rural/less profitable routes. Most of the passenger airports in the US for instance would not be viable without subsidies (most flights go to a few profitable hubs, but the long tail or airports forms the majority by count).
Amazon delivers everywhere because USPS subsidizes package delivery to unprofitable areas. You don’t get next day prime except in a relatively small proportion of the country (by area).
What is your point? Sometimes it works due to other factors so we don't need regulation?
Beyond the very real closures of unprofitable rail lines that is always happening or threatening (including in state owned rail companies) your examples don't really measure up either:
Local airports are typically subsidized. So the infrastructure aspect is not market driven.
Eventually getting starlink means it was fine to not have fast internet for a couple of decades? We still don't have good mobile internet on many train lines in Germany. In Poland it's fine. The difference? Poland mandated that companies bidding for 4G meet certain coverage goals. Market forces have absolutely failed to actually get reasonable mobile internet to everyone.
I think when you say “profit motivated” the underlying principle is actually utilitarianism; doing the most good for the most people, for which profit Is merely an imperfect proxy.
“profit motivated” means taking as much as possible while giving as little as possible in return. The ideal situation for a profit motivated company would be collecting all of the money people have and delivering nothing. Thieves are also profit motivated.
This might be a cynical take, but I doubt Ticketmaster (and most of these other examples) are motivated by doing the most good. Their underlying principle is extracting the most value for shareholders at any cost.
Some people argue even that behavior ends up producing the most good, but I cannot accept that level of mental gymnastics.
There is a limited amount of engineering resources. Don’t you agree they should try to use them for the benefit of the most people? That’s not to say they should exclude groups. But a small minority path probably gets less work than a main path. Would you agree these prudent or not?
Yea, when this topic comes up on HN, a lot of the usual excuses appear: It's hard to write software that works everywhere! It takes too long to test on more than one browser! It's too expensive to hire someone to port to X platform! We're trying to bootstrap in a hurry--there's no time to support Y people! Everybody should just upgrade to the latest, why should we test on older systems?
These are attitudes come from the privilege of never having been in that 2% of users, and I think we have them until that one day we end up being in that 2% and can't use the system ourselves.
When I wrote iOS apps, I was constantly infuriated by the tech lead's and product management's insistence to only support the current major OS version and the previous one. Engineers would take time out of their day to rip out support for iOS X-2 (rather than fixing bugs, working on performance or features)! Code that wasn't in the way of refactors, wasn't really buggy, wasn't harming anything architecturally. To me, it just looked like Griefing The User. I didn't get it and I still don't. Now, I have a 8 year old phone, and lo and behold, half of the apps in the AppStore don't even work on it anymore because of this attitude, so I guess I'm firmly in the 2%.
Hey HN, we're building a tool aimed at organizations who use a lot of different AI assets (skills, tools, rules, agent definitions etc.) and find it hard to manage them consistently.
If anyone is willing to give us a bit of feedback, I guess we'd mostly like to hear if the issues this blog post opens with are even a problem for you and if yes, how are you currently solving it.
reply