Posted by stalfosknight 1 hour ago
I'm sure that right at this moment at least some people are thinking "if only we had a different executive, then we could rein in this AI problem." That is wrong at best. You could rein it in for ~4 years until you lost the next election. With a completely feckless Congress, very little can get done.
But the admin has repeatedly ignored such restrictions. This check on power also loses it's teeth when the oligarchs align themselves behind the executive branch.
- impeach
- pass new laws / revise old laws
- hold real hearings
- amend the constitution
Now practically will this congress do so? No, but in principle they wield more power than the executive.
Of the dollar, by the dollar, for the dollar ...
Honestly, calling Congress castrated is fine because it is healthy venting about how ineffective they seem at their charter.
Even though it's really charming and compelling to believe, there is no one solvable problem simply requiring elbow grease, voting harder, proprioception of comically vulnerable reproductive organs, etc.
Members of Congress, just like everyone else, act in their own self-interest. And unfortunately for pretty much everyone else, their best method of self-preservation is to do nothing, hence the "eunuch" Congress.
It's like if someone is building a landfill in your hometown to bury the whole country's waste. Or it's like a factory that creates zero job.
This is a well established playbook - it was used with nuclear. It’s being used with oil transport.
It’s literally the same script.
In the worst case, if your local municipality sides with business over the little guy, that means potential brownouts and water shortages for you.
Key point doing a lot of heavy lifting here. Do all these data center buildouts include providing their own power? Seems like the answer is largely no. These companies expect power infrastructure to be supplied by the government, but also want lower taxes.
Taking a step back, if the US unilaterally stops producing AI will other countries stop? The answer is clearly no.
Datacenters and ai can be built and trained anywhere. If you want control over AI you should want it to be built in your own country where you have political representation.
All preventing datacenter buildout will do is ensure that the price remains high and only really rich organizations can access it.
If you look around, you'll find the AI has made some significant improvements to medicine and engineering. These improvements get drowned out by the AI Cheerleaders, but they're there.
I keep seeing this and I'm pretty envious! You must have a different form of ADHD than I do. For me, trying to use AI to build anything is terrible for my attention, it turns everything into a miserable slog because it's so hands off.
I miss getting into flow.
Despite this, because of its other effects, the cost to clean up and stop using asbestos is greater than the sum total of any benefit from all mined asbestos worldwide.
Even a miracle technology can still be a net disaster.
Now you understand why anti-vaxxers ignore evidence. Because it doesn't fit with your worldview and you're too narrow-minded and selfish to consider that your viewpoint might actually be wrong and bad for others.
They don't provide appreciable community value and they effectively mine limited local resources (power, grid capacity, land, water) and sell it as compute, immediately diverting the profit back out of the local economy and into very distant business accounts instead.
Builders choose their targets specifically by how well they can strong-arm weak/vulnerable communities into letting them build these mines through political influence and misrepresentation. It's bad.
What you probably want is to leverage their global market value to establish new power and grid capacity in undeveloped areas, perhaps to someday become a seed for new communities that grow around the infrastructure development work.
But that's much more expensive than bullying and seducing a weak city council so it won't happen with regional/state/federal regulatory protections or incentives that push them away from the exploitative opportunities and towards the constructive ones.
The greatest control I have is probably to have it not get built, though even that is minimal as it has failed to stop the one that is indeed being built in my town.
So urgent that Andreesen has a Super PAC to push the dangerous China narrative.
It's easier for normal people to influence local regulations, but local regulations just push the problem somewhere else. However disdain for AI is so widespread that this is actually kind of effective.
I think it's a mistake to stop producing those things.
Is the reason that competence and social trust are declining because of AI? Maybe, but not only that.
The fight against AI should just be about taxing token usage. We should also tax the hell out of anyone using AI as an excuse for layoffs. It's far past time to ban buybacks and dividends for any company doing layoffs. We also should have a requirement, you have to provide a bonus pool that goes dollar-for-dollar for any buybacks or dividends you do.
There’s no such thing as a red state or blue state, these are fictions created to generate political fighting for no value to society.
Second - many states such as Ohio have begun pushing back strongly against data centers. In Ohio we had been offering tax breaks for construction because we welcomed the economic activity, but thankfully the government here after seeing a lot of pushback across the state has realized providing tax incentives or subsidies is economically and politically stupid relative the benefits of the new data centers.
To your point, they can be built anywhere. So many folks are saying yep, let’s build them somewhere else and drain water and raise energy prices there instead of here.
Smart politics in a state like Ohio would require data centers to relocate corporate jobs to the state or face full or perhaps even surcharges for utility rates because why not?
Not again...
this is especially true for AI use cases, where compute is hugely more important than latency / bandwidth
> you have to provide a bonus pool that goes dollar-for-dollar for any buybacks or dividends you do.
So, reallocate some exec comp to a pool that gets bigger when you give shareholders back money?
Would be great to balance the market better between labor and capital, but there's no easy button...
What about self-hosted models?
If it was AI they would take those extra people to get more done. I know of no company that doesn't have more work than they have people. (but they lack the funds/ROI to pay more people)
I mean if AI is really powerful the reason is "profitability as our competitor steals our contracts at a fraction of the price". Your competitor just doesn't hire in the first place of course.
(Until the bots are still more efficient)
Instead of a ban just make sure they pay what's needed to keep capacity where it needs to be.
If I build a house on undeveloped land and the electric company needs to run lines, do I also (in a much smaller way than a data center) increase the costs for all other customers? Is everything always just spread evenly?
Energy hungry infrastructure projects pay something called a "large-load tariff" to try to contain their second-order costs from leaking into residential rate payers pay for. It's not perfect, so a datacenter project could trigger some upgrades that cause rates to go up.
The situation is confusing everyone right now because it's impossible for the average person to tell why rates are going up. A lot of utilities are doing things like finally addressing old fire-prone infrastructure (see the California fires) and dealing with inflation for everything from their generation input costs to inflated costs for infrastructure to putting straight of Hormuz-inflated gas in the tanks of their fleet. Customers only see that their rates are going up and AI datacenters are on the news, so they put them together and assume datacenters are to blame for everything. Yet rates are spiking even in places with zero datacenters.
The topic has entered the domain of emotionally charged topics so nuance is hard to come by. Many of the anti-datacenter people are against datacenters as a proxy for their hatred of AI and the electricity and water arguments are just convenient justifications. This is how we arrive at the article.
electrical supply is not infinite. datacenters have high electrical demand. more demand + same supply = increase prices.
>Why doesn't the data center bear an extra cost for the added infrastructure?
the problem is that added infrastructure is not built instantaneously. it lags behind. so costs will be high until more supply-side infrastructure is in place.
i agree that there should be some sort of stipulation that when you build your mega datacenter that you also have to build out electrical infrastructure at the same time. but unfortunately, that is not how it is.
I totally get the general principle that not everything scales linearly like that. But, I also know very little about electricity generation, so I have no idea where the breakpoints are. (I would also guess that if the demand dips low enough, there could be a case where after decreasing costs for a while, costs actually start to rise again as there is some minimum infrastructure needed but fewer customers to bear the cost.)
I suppose you can reduce a lot of both good at bad things about the country to "because federalism".
In some states (like Oregon and Virginia) they do, but in a lot of states the regulations for rate structures are flat among all users so when there's a large surge in new demand the utility will build out new capacity and spread the cost of that new capacity to all rate payers with no regard for the fact that the new capacity would not have been needed without the new demand (from data centers). So everyone who was already using the electricity pays the new higher rates along with the new large-load user.
These companies building data centers will often make a lot of PR statements about how they are fine paying the extra cost for extra use while at the same time lobbying behind the scenes to actually avoid that happening and fighting against changes to utility rate structures that would raise their costs. By and large they can't be trusted.
The cities are paying exorbitant prices for electricity to pay for safer infrastructure for rural customers (undergrounding).
Some cities have divested from PG&E and enjoy much lower electricity prices as a result.
We're dealing with this bullshit in my own city in another state. From the parks to the roads to the sidewalks to the library every goddamn thing we touch gets driven up to the point of "can't actually do what we wanted" in cost because some rich assholes 100mi away in the vicinity of the capitol have taken a "build it fancy and rich or don't build it at all" attitude and enshrined that in state law and rules.
Sometimes they'll be so kind as to eat part of the cost with state grants, as long as we sell our freedom away in other ways.
In some sense, sure, any time you buy something you apply some upward price pressure. Of course, whether the resulting price changes measurably depends on many things, like the scale of your purchase relative to the scale of the market, the price elasticity of demand, who the marginal buyers and producers are, etc.
First, increased demand drives increased prices. This is the least controversial axiom of modern economic theory. So if you add a huge power consumer to a market, all consumers in that market will have to pay more. You can mitigate that some if that new, big consumer builds their own power facility, but the fact still remains that the local price in fuel (oil, coal, etc) or materials for renewable generators (turbins, solar panels, etc) will increase. Again, because demand increased.
Second, its one thing for things to cost more in a market that has a booming economy and plenty of high paying jobs. Home prices in the Bay Area are horrifying, but the poverty line for a family of 4 is $80k, which sort of grounds things. If energy costs go up by $100/year in the Bay Area, nobody notices. But if energy costs suddenly skyrocket in Great Falls, Montana (poverty line for 4: $33k) or similar that lacks a vibrant economy, the residents don't have much choice but to tighten their belts over the suddenly larger electric bill that has done basically nothing to actually revitalize their economy.
This is a wall of text but genuinely worth skimming: https://www.wired.com/story/us-law-enforcement-warns-of-anti...
The past 20 years of surveillance capitalism and the general deployment of technology against consumers should make everyone question whether this could ever be possible.
I know it because i tried ...
1. Electricity costs in Maryland jumped 89% over the past year, much more than anywhere else, largely due to an AWS data center expansion: https://www.visualcapitalist.com/mapped-where-electricity-pr...
2. At their heart, data centers are extractive. Their boosters always overstate the jobs they will create, but they basically take land and resources from one place and create the vast majority of the wealth somewhere else. They are giant windowless boxes, they don't support their community in any way, and in fact with AI they basically add to more job destruction in their communities.
While I agree that some downsides of AI are overstated (like water usage), this whole article smacks of paternalistic "the peons just don't understand what's really going on" nonsense. The same thing happened in the 80s, 90s and early 00s when many economists painted those who lost their jobs due to globalization as Luddites who just didn't understand economics. Only decades later did many economists readily admit many of the huge downsides to many populations from globalization and that reskilling rarely works.
The question is whether globalization is a net positive and whether people understand that, even if it comes at a cost to themselves personally.
It should be noted that modern globalization (post 1945), with Bretton Woods, GATT/WTO, container shipping, and eventually the internet creating the integrated global economy we have today, is but the latest milestone in a long history.
Industrial wave of 1829 - 1914 radically reduced the cost of moving goods and information.
The first globalization happened 1490s - 1800s, when Columbus’s 1492 voyage and Vasco da Gama’s 1498 route to India created the first truly intercontinental trade and migration networks, linking the Americas, Africa, Asia, and Europe for the first time.
That's moving the goalposts. While economists acknowledged there would be some disruptions, in the 90s the vast majority of them downplayed what turned out to be prolonged negative effects to huge populations in richer Western countries. And, ironically, it was those negative effects that led to the rise of nationalism and the dismantling of globalization that we see today. I understand folks' objections to some of the opinions and writings of Paul Krugman, but I give him credit for admitting he was wrong: https://www.bloomberg.com/opinion/articles/2019-10-10/inequa...
Sure, there aren't many people who deny that globalization lifted millions out of poverty in East Asia. That's not really comforting to folks in the Rust Belt whose communities have been hollowed out and devastated.
Blocking them should be a priority until rates are negotiated with your G&T / major provider (PJM and FERC in Maryland and many other states)
RE blaming the peon reader: you’re talking about Vox so that is expected unfortunately.
Best part is when they figure out how to take their $10-20k donations to the local community as "doing good" and turn it into positive PR spin to the local yokels that don't know any better that they're getting robbed blind daily. They have a full playbook to rob local communities and get them to fall for it and it sucks.
I've watched it happen in Upstate NY and we certainly aren't seeing any benefits of any of it.
But yes this article is absolutely the "usual sort" of paternalistic garbage.