51–60 of 151 posts
At the same time, the more the US government acts to support data centers, the greater the reaction will be against them. It is the classic situation where the more the government does, the less people will like it.
It seems like there will be a collapse in building data centers, to the point that a large amount of the newly built ones will be reclaimed for other things. There is a very large group of people who are against building new data centers because they hate AI (and any pollution related arguments these people make are in support of their primary concern, being AI). But on the flip side of that, even people that I know…
We'd originally set this up to be able to locally run larger models, in the 300-400 billion parameter range, but the rate of improvement of open-weight models has been so fast that, coupled with extensive custom skill creation, we're now getting similar results out of Qwen3.8 27B to what we were getting out of Qwen 3.5 297B when we started out with the project, which frees enough memory to allow 20-25 users to have 256K context concurrently. Both the hardware, the software, and the models are improving at an accelerating rate.
Investing in data centers to support SaaS LLM providers today feels a bit like investing in mainframes and minicomputers in the late '70s, with a massive paradigm shift lurking right around the corner.
Actually, it's probably already closer to the early '80s, given that purpose-built local AI workstations are already available at price points lower than the inflation-adjusted initial price of the original IBM PC.
What I don’t understand is that the vast majority of LLM workloads are pretty damn latency-insensitive. Why can’t they build these data centers where there’s little to no opposition? Sure someone will need to maintain the data centers, but with trillions in investment lined up, surely they can just pay people enough so they’re okay with living in the middle of nowhere. Edit: When I say the middle of nowhere I mean sa…
It's difficult getting a grid connection and fiber optic uplinks in the middle of nowhere.
What I don’t understand is that the vast majority of LLM workloads are pretty damn latency-insensitive. Why can’t they build these data centers where there’s little to no opposition? Sure someone will need to maintain the data centers, but with trillions in investment lined up, surely they can just pay people enough so they’re okay with living in the middle of nowhere. Edit: When I say the middle of nowhere I mean sa…
Datacenters need a source of water, and human populations are usually concentrated in those places too for the same reason.
Earlier quoted context omitted.
I'm sure the luddites got all their ideas from TikTok aye? At least were not at this point yet: "Mill and factory owners took to shooting protesters and eventually the Luddites were suppressed by legal and military force, which included execution and penal transportation of accused and convicted Luddites." https://en.wikipedia.org/wiki/Luddites
"The luddites were right" seems like an odd position to have on HN.
Earlier quoted context omitted.
I'm sure the luddites got all their ideas from TikTok aye? At least were not at this point yet: "Mill and factory owners took to shooting protesters and eventually the Luddites were suppressed by legal and military force, which included execution and penal transportation of accused and convicted Luddites." https://en.wikipedia.org/wiki/Luddites
"The luddites were right" seems like an odd position to have on HN.
Earlier quoted context omitted.
FTA: 'The largest estimate for any single facility ever.' From the researcher: 'We estimated between $53 million and $99 million in annual health damages from associated air pollution — among the largest health-damage estimates produced for a single facility to date.' [0] Still enormous, but there is a significant difference between being 'THE X of all time, and being 'among the X of all time'. Regarding 'minor sourc…
And this reasoning is exactly how we got here in the first place. “It’s not that bad, just a little bad, so it’s okay” We’re talking about air pollution and death rates, not financial risk.
> The latest proposal applies to permits for a "minor source" of air pollution ... [the Vantage data center was estimated to contribute] between $53 million and $99 million in annual health damages, the largest estimate for any single facility ever. 'The largest estimate for any single facility ever.' 'Minor source.' I give you double-speak, gentlefolks of HN. Name it something - anything - and you frame it in those…
FTA: 'The largest estimate for any single facility ever.' From the researcher: 'We estimated between $53 million and $99 million in annual health damages from associated air pollution — among the largest health-damage estimates produced for a single facility to date.' [0] Still enormous, but there is a significant difference between being 'THE X of all time, and being 'among the X of all time'. Regarding 'minor sourc…
A minor source should never lead to deaths or to put it otherwise No deaths = minor source
deaths = major source
Many deaths = majormost source
Now we just need define many but we already know its a major source of pollution that kills people
It seems like there will be a collapse in building data centers, to the point that a large amount of the newly built ones will be reclaimed for other things. There is a very large group of people who are against building new data centers because they hate AI (and any pollution related arguments these people make are in support of their primary concern, being AI). But on the flip side of that, even people that I know…
How is anti pollution argument pro ai? And even if you do inference at home you are not training the models. Moreover, most people certainly are not doing inference locally.
What? (Edit: I think you misread "in support of their primary concern, being AI" as "in support of AI"? People who complain about pollution are usually mostly concerned about AI, and they use arguments of pollution to strengthen their argument against AI related things.)
> And even if you do inference at home you are not training the models.
No, and training does use a large amount of energy on a large amount of hardware. But:
- That sort of workload doesn't really require many distributed data centers, only a few powerful ones. I believe training is also getting cheaper for the achieving higher levels of capabilities, but I don't think the efficiency advancement has been as dramatic as it has been for inference.
- I believe we are really hitting a wall of diminishing returns, especially at the high end of large models. There is lower demand for training, because models are good enough to have a reasonably long shelf life at this point. Year+ old models that were SOTA in their time are still useful today. The demand for training is going down.
> Moreover, most people certainly are not doing inference locally.
Maybe not, but I think most people who go out of their way to use LLMs because they find them useful for their work actually are. Though most inference is probably from people who do it accidentally (as a part of a search result or something) or students who don't have the resources to do it themselves. But basically every software engineer that I personally know that uses AI for programming is either running their inference on their own hardware, or is talking about building a rig to do it.
This degrowth anti-technology mindset just does not make sense to me, and I think it's mostly astroturfed from TikTok.
- no NDAs signed with elected officials who then stonewall the public and hide behind them - paying the same prices for water and electricity that I pay - paying the same property taxes and income taxes that I pay - building in a fashion that doesn’t leave area residents stuck with bigger electric bills and water bills
All of these things can be done.
I have a tiny data centre. I get no sweetheart deals on the above tho bf s.