because no one believes there are legal consequences if they don't
and there are a lot of ways to doge it even if there where a reliable government in place
like especially if they do what they have been doing recently (run their own generator, build their own power planes) a lot of this cost is implicit and as such very dogeable. E.g. higher cost for gas power planes for other due to major increase of demand, higher medical cost due to more air pollution, higher fuel prices, etc. etc. (not even speaking about anything climate change).
If it is not legally required, it will not be done.
Blinded by greed, I have never done it before, but I have seen the light, the bright future that we are all building toward.
From this day forth, I shall be righteous.
In your name, all good things come.
Hallelujah.
There's a lot of data centers that are not being built because they're not fixing it. The trend is going to continue. The hate for AI is going to grow. You basically have a lot of people that will vote a lot of people into office to take down all AI progress inside the United States if they don't fix their problems.
It'll be cool to shit on big tech as a politician.
And it's not just data centers, it's all sorts of industry. My local gravel and concrete plants run their "big stuff" off generators because the cost of the utility drop for their amperage doesn't make sense. And nobody will connect the dots between these choices and the requirements we've saddled utilities with. They're spinning up generator not because it's cheaper per watt, but because they're not operating on the 40yr timeline you need to be in order for the red tape you have to go through to put in permanent infrastructure to pencil out.
I'm an abutter for a utility project and I've gone to the meetings for and it's an absolute massive boondoggle. My energy bill is going to reflect god knows how many hundreds of billable hours it takes for these hired lawyers and engineers to prove to the system that they're not gonna fuck over any endangered frogs by widening the cut to meet some industry standard that changed over the past N year and dumping culverts and fill in some places where streams criss cross it.
Literally nobody involved cares. The abutters don't care. The town wants it to go forward because it's all trivial and it's not like it won't be their ass if they block an upgrade to industry standards and something happens. The system is just going through the motions. The city engineer grills them about petty bullshit because it's literally his job. They know he will and they have the answers but he makes a show out of the subjective things. Ditto for the conservation commissioner. It's like the Israel missiles meme. One side is my tax dollars and the other side is my energy bill. We're all doing this because some slimy politicians wanted to pander to some shortsighted big picture ignoring environmentalists 50yr ago and beurocacy has perpetuated and grown itself since. No public interest is served by this.
And the cherry on top is that at the margin, we get shit like generators that don't need to exist because the cost of the alternative is driven up to the point the fuel inefficient (and also dirty) solution makes sense.
I'm more than insinuating it. The cause may be noble but the state and local implementations have been perverted by all the minutia and business interests and NIMBYism. I would go so far as to say that (state and local implementation of requirements within) the clean water act is a non-negligible contributor to the decline of manufacturing and agriculture in the northeast and upper midwest.
We basically took "thou shalt not dump for that is bad for our surface and ground water" and over 50yr turned it into a blank check for all manner of leeches to make a buck and all manner of NIMBYs to make things unnecessarily expensive.
Petty 1.5-acre "I want my lightly forested former field to be a field again" and "business is going great I want my gavel parking lot paved" being stalled by five figure costs and even with those costs incurred it doesn't guarantee compliance. That's a far cry from the "yeah we just dump this stuff in a settling pond, IDK where it goes after that but man that river over there sure is a weird color" type 1960s industry behavior that it was meant to really curb. And the big industrial offenders still get to do what they want, not as bad as before of course, but still bad. Some Megacorp's runoff might turn the fish neon green or their 1k unit condo development might turn the river brown with silt but of course they'll be right there with their lawyers and experts who'll tell you why it's fine even when it's not to anyone with a brain and two eyeballs. The layman can't pay off people like that be on their side and neither can the regulators. (I assume this frequent fact pattern what you refer to by "living in the stink")
I think it ought to be revamped at the state and local level into something that's substantially more "results based" rather than the proactive red tape "make the bureaucrats feel like their ass is covered if someone ever complains about what they approved" based system we have now.
(And just to be clear for any readers who aren't familiar, the clean water act basically doesn't do much to affect the average person or business at a federal level. The local implementations and all the key definitions, industry standards, etc, are where the rubber really meets the road)
Edit: Basically I'm saying that in the past 50yr the interests the CWA was supposed to stymie learned how to pay their way around it, the parties who make a buck doing that have gotten themselves all but written into the compliance process to the detriment of the interests the CWA was supposed to not seriously burden. It needs to be replaced or revised to solve those two big problems.
Extinction is forever -- your frustration, and that committee process, does not compare directly to species extinction.
as long as the boxes are checked, most public sector employees are not going to stick their neck out. it's steady work that pays all right and has great benefits (and even pensions sometimes). a lot of people in the public sector aren't willing to step out of line for "frogs" even when they should or want to.
So you get some guy approving some megacorps project because "well it says here that they've tested the stuff and the liquid mercury is below the <cites EPA mumbo jumbo" and giving farmer Johnson the runaround on some land clearing because farmer Johnson's project isn't worth spending $30k on civil engineering to have someone use "the right words" to say "I'm not gonna muddy up the river" and while the guy in the office knows what farmer Johnson means, he can't approve the project without sticking his neck out and so he doesn't.
It must be nice to live in a desert state where every 3rd parcel isn't "within regulatory range" of "wetlands" (most people don't know how expansive that word is).
Something something the more numerous the laws...
I used to think we were progressing up an exciting tech tree. That seems naive now.
Water, land, energy, the soundscape, intellectual property that incentivizes the dissemination of good ideas, digital networks of information and self-expression, perhaps even the economic value of expertise itself are all being sacrificed in the now for promises of utopia in the future.
Precious eggs to give to those promising a utopian omelet, eventually.
Exactly opposite will happen. Reason is, when Big Tech is paying huge amounts of money to contractors to build those power generation facilities and service companies to service it, they will abandon servicing other facilities (remember how Micron dropped consumer RAMs last year because of enterprise demand) or require higher pay from everyone else
1. DCs must be built anyway
2. You can't take away energy from households
(3). Highly preferred that you are not going to impact cost negatively to households (otherwise why we have this discussion)
based on these assumptions, solution I see is, BigTech subsidising energy costs for 10 years for nearby households (area will be geofenced, e.g. in the radius of 50km), subsidy will be based on the prices outside of that radius. e.g. if you everyone outside of closest DC pays 1$ and in the radius prices become 1.5$, 0.5$ will be covered by BigTech and they're also responsible & pay to setup the system to automatically include everyone in subsidy program, not like you need to apply
Also BigTech is not going to build the power generation plants, it must be built by existing processes to minimize impact on pricing
* power generation is not local in most of the cases. You'd be still fucking energy market to anyone outside it
* power is also used by companies people actually want. Even if household power cost wont chance thanks to that approach, the price for power for every single business around will increase.
* similarly any other manufacturing business will cost more. In essence, the AI boom will reduce profit margin of every single business that has electric power as significant cost in the production. Which is a lot of industries.
"Making AI pay more for its useless power draw" is nice idea but it is pretty hard to realise. Unless we start outright denying connection to power grid but that's pretty dangerous political precedent to set.
And should never be a thing. It's not for others to decide something is useless. Let the market figure that out. It should simply be an even playing field, and in theory we already have the regulatory mechanisms to get that done. If the rates are currently able to be abused, that is entirely correctable.
It should not matter if I want to build 1000MW of aluminum smelting capacity, or a facility that digs holes and fills them back in again - so long as I'm paying the regulatory rates that have been set to be fair across all similar power consumers.
The bigger problem by far is that we spent 40-60 years not building anything. Eventually you run out of the previous generation's power capacity. Can only ride inertia and offshoring your industrial energy usage for so long. There are a lot of chickens coming home to roost in the next decade or three on this topic, electric grid was just pulled forward a bit due to the AI bubble.
It's the entire region where prices will increase. A facility could be built 200 miles away from you and it would have just as much of an impact on your power bill as one built across the road.
> Also BigTech is not going to build the power generation plants, it must be built by existing processes to minimize impact on pricing
The issue with this requirement is that it's not a capital problem. Big tech would be happy to pay someone to build more power generation and just take a power feed, even if it costs marginally more.
What they cannot do is wait 12 years for what they could get done in 2. And I do not blame them at all for this at all - building anything in the US is basically impossible these days due to the NIMBY and general culture of not getting things done. We forgot how to rapidly build infrastructure in an efficient manner, and anyone putting proverbial boot to ass to get this done is a good thing.
It's not a case of a datacenter operator wanting to save a few pennies per kwh. It's a case of the power literally not being there and the utilities being utterly incompetent at building anything in any reasonable timeframe.
There's some speculation in the comments about what is or isn't in the pledge. I recommend reading it yourself.
[0] https://www.whitehouse.gov/articles/2026/03/ratepayer-protec...
[1] https://www.whitehouse.gov/presidential-actions/2026/03/rate...
> IN WITNESS WHEREOF, I have hereunto set my hand this fourth day of March, in the year of our Lord two thousand twenty-six, and of the Independence of the United States of America the two hundred and fiftieth.
One can do the former whilst repudiating the latter and remain logically consistent.
We just approved the first nuclear plant in 20 years to a company owned by Bill Gates and in a state that has basically nothing but farmland and a Microsoft datacenter.
This absolutely cannot backfire. /s
The threat is: This "datacenter power" disincentives buildout of "free" powerplants (by eating up significant demand at very low margins thanks to basically vertical integration); this slows down buildout of "normal" infrastructure (possibly both grid connectivity and power), and the electrical energy market becomes worse for consumers than it is now.
I personally think all of this is very speculative for now, but allowing industry to rely on the grid (which they still would!) while almost exclusively "buying" their own power is a risky proposition from a consumer perspective.
Interconnection expenses.
Same issues as with mining and large industrial clients generally.
In a different world they would have earned trust and deserve the benefit of the doubt. This is not that world.
In short, I'm very much in favor of building the right solution to a problem.
I am unsure what cognitively triggered an unhealthy response of "this is NIMBYism!" and would welcome a follow up comment to understand your train of thought.
It reduces profit.
Almost everything in tech has been subsidized in one way or another via tax avoidance schemes or outright lobbying and manipulation of the market.
Why would this be any different?
Increasing natural gas generation is of course disastrous policy with a major death toll from the climate disaster, there needs to be a rampdown of fossils use and production.
So no.
But what probably also isn't included but should is environmental damage.
Running low quality "temp." gas turbines non stop isn't without filters etc. isn't just bad for the climate, it's a air pollution which can directly affect anyone in it's path with not only increased chances for lounge cancer but also much more short term effects like asthma, and increased chances of asthma attacks ending deadly. Especially if the weather prevents easy dispersion (like it tends to do in winter). It's not that long ago (<80y) that the west had acid rains, and deadly smog accidents exactly from this kind of negligent shit. And if we look at Asia this is sometimes still a topic today (but has gotten much better compared to just ~20 years ago).
No MBA pencil pusher wants to run an inefficient local turbine. It's just that the timeline and upper cost bound of doing that is less crap than having a "real utility" build more power at "real utility scale" and run you a wire because the latter is subject to all manner of delay and cost overrun.
And there's no inherent physical or economic reason for it to be that way. We made it that way. The metaphorical local turbine is less worse specifically because people like you, saying the exact same things you're saying right now have saddled the "real utility scale" generation, and more importantly, the wire to the big industrial consumer who'd pay for it with all sorts of requirements.
It costs tens of thousands of dollars of lawyers and engineering over years just to dump a concrete culvert in a ravine where it crosses a power line clearing and fill over the top, all because of the red tape. Say nothing of the cost to do all the legal paperwork to get the utility cut in the first place. Now multiply by every mile the wire has to go, add in the wires, etc, etc. For an industry that might boom and bust in 2, 5, 10yr dumping a fuel guzzling turbine in your parking lot at 5x the cost per watt starts to look pretty good.
Of course this will all change, but I doubt we will see tech companies opening power plants anytime soon with their associated balance sheet dragging 1% return on equity.
When our civilisation is excavated in 500 years, they are going to say we were as crazy as all of the others.
trees aren't just carbon, they are bio mass/nutrition
and if you constantly remove bio mass you sooner or later run into issues
(Which we already do in some places, e.g. when over using fields (see US dust storms), or with some managed Forrest getting increasingly more unstable not just because of warmed climate but also because of removing dead treas leading to an interruption of the natural nutrient recycling (and insect habitats) leading to Nutrition deficiency in the long run.)
but we do have working carbon removal technologies, they are just not cheap
hence why you want companies to pay for them, it gives them a huge reason to reduce emissions instead
The USA currently produces about 70 million tons of paper per year, which is about half carbon by weight. We produce about 2 gigatons of lumber per year, which is again about half carbon, all absorbed from the atmosphere.
Unfortunately, we produce like 40 gigatons of CO2 per year. So we would have to scale lumber work dramatically. It's also not a clean industry itself, reliant on heavy machinery running on gasoline or diesel, and turning that wood into charcoal would require massive refineries.
IMO more effective bets are figuring out how to artificially induce massive blooms of algae and plankton in parts of the ocean to essentially recreate the conditions that lead to the hydrocarbon deposits in the first place. There's some work on this right now, but like any massive engineering and ecological tampering, there will be tradeoffs and downsides. I also don't know how you prevent the dead plant matter from decomposing and releasing the carbon.
completely ignoring all existing technologies related to that topic to spout obvious nonsense about "cutting down trees and burying them" (which would bind active bio mass which isn't a grate idea, also that won't produce oil anyway not that this is relevant for the discussion)
various ways to reduce the carbon in the air do exist (and without trees)
and the carbon can be both recycled for other usage and literally placed in the earth, too
it is not rally a solution for climate change as it's very expensive to do. But this also makes it a good idea to "make companies pay for it" (at least if their carbon-equivalent output goes above a certain threshold). Because if they have the choice between very expensive carbon removal or reducing carbon output for a much cheaper price they will do the later; But in emergency/outlier situations they still can do the former, just at a very high price.).
Theoretically you could harness some of those volatiles for some energy production, but at the very least use those volatiles to heat the wood and make it charcoal for basically free.
Actually, the tweet quoted in the article is firmly in the "you can't make this $%&/ up" category...
It is also a ~50 year investment.
This makes it not very attractive for companies and is why most nuclear power is state sub-ventioned.
Theoretically the US had something similar to a state bank to help companies to finance exactly such projects, but Trump/DOGE defounded it for publicity reasons which makes it even less likely for private nuclear power plants.
Many "we will use nuclear power" statements do rely on mini reactors. But AFIK pretty much all mini reactor projects have ended in dead ends so far. With promised at best working out on paper (and quite often not even there).
So my guess is: They will claim they want to use Nuclear and might even intend to do so. But in the end look at their balance sheets and risk calculation and go "nah, lets do coal/gas/oil". There probably will be some single public co-investment into a nuclear power plant which "happens" to also be government sponsored to keep up the pretense.
Ta da! It's 3 Megapaks wired together with a transformer. Now that's innovation.
See? Without billionaires, who would save us from billionaires?
Tesla’s Megablock: Breaking the Transformer Bottleneck https://www.rebellionaire.com/post/tesla-megablock-transform...
Building pipelines is a long and arduous process and one that will not be done in time to reasonably accommodate the increase in natural gas demand presented by these data centers.
Soon I’ll get a used EV and cover the garage in panels too so I don’t have to care about wars causing surges in gas prices either.
- They already invest in new power plants and connection infrastructure when they bring in new datacenters - Electricity for datacenters is based on capacity rather than actual usage - They already have backup generators at most datacenters that they can run during outages. It wouldn't be much work to allow those to feed power back into the grid in extraordinary circumstances - They generally use local contractors to build them for practicality purposes anyway.
This is just some fancy PR and nothing else.
Launched in 2010 by Bill Gates, Melinda French Gates, and Warren Buffett, it was sold as a historic shift in philanthropy. Fast forward to 2026, and the data suggests it’s been more of a "Wealth Preservation Society" than a massive wealth redistribution event.
This will be just as trustworthy. We need laws - not merely rhetoric pledges !
The pledge is meaningless political theater meant to placate voters in the mid-terms.
One could say "use renewables", but even that has externalities: that means an increase in demand for solar panels, or wind turbines, or the labour to maintain them, which again leads to an increase in prices.
I'm sure we'll be hearing all about how much this benefits households in the coming months and years.
That's why he wants to go into space (10x solar potential because you don't have a day/night cycle, no clouds, no dust/rain, no temperature loss, no orientation issues, and no atmosphere reducing solar).
To me it seems ridiculous, for one because sending 150kg to space costs about $500k, and this is about the weight of a solar installation that costs $800 to install and generates about $1000 worth of electricity across 20 years at utility wholesale prices.
But suppose it was cheaper and viable, and earth-electricity was indeed capped, you could argue (if you believe the hype) that developing AI is an existential arms-race objective for US/China.
But from what I've understood that's just not the case at all. Something like 170+ coal plants are scheduled to be decommissioned, and the average coal and gas plant runs at 40-50% of capacity, because wind/solar is eating their lunch (cheaper marginal $ per kWh). i.e. there is so little demand that these plants keep using less capacity and shutting down superfluous plants.
You'd think if experts believed electricity was going to be a bottleneck, that venture capital / AI companies, or even traditional capital, would be buying up plants or signing guaranteed-usage contracts. But it doesn't seem to be the case.
In fact, Nadella publicly stated that he has a large amount of hardware in inventory that has already been purchased but cannot be utilized due to insufficient energy.
The general goal for utilities has been to pursue the next “thing” and work toward some sort of regulation to lock in demand, which can be used as a lever to seek price increases and consolidate.
If there’s margin to be had, the utilities will find a way, and prices will go up either way.
Pledge my ass. It is either law mandating those massive datacenters absorb the cost with heavy penalties for non compliance or it is just BS talk (what it seems to be at the moment)
I can see how big tech is enthusiastic about freestyling this. Eh sorry I mean bear the cost
Take the case of Duke Energy in North Carolina, which illegally raised rates too much. Utilities prices are supposedly regulated but utilities work around this by simply moving costs to things they can charge whatever for (eg transmission costs vs energy costs).
The NC Court of Appeals ruled that Duke Energy's actions were illegal BUT there would be no refunds for customers [1], in part because lawmakers passed a law to allow them to do this retroactively [2]. Also, if Duke Energy had to repay customers they can simply raise prices to recoup those costs even though the money was improperly charged in the first place.
So consumers will keep paying for the infrastructure to connect up these data centers and will keep subsidizing the ongoing energy costs.
[1]: https://www.wcnc.com/article/news/local/no-refunds-for-duke-...
[2]: https://sustaincharlotte.org/press-release-nc-lawmakers-over...
People are not voluntarily going to build things that make less profit.
It is a suckers bet assuming the unscrupulous will grow a conscience. =3
Saying they’re going to pay for generation and transmission adds little. That’s already baked into the charges! It’s like saying they’re going to finally pay for the farmers to grow the produce and the drivers to get the produce to market when they buy apples--as though spontaneous generation and teleportation was ever an option.
They'd ask the utilities to make Gigawatts of energy available over the next two decades and the utilities would say "No problem, just sign here and agree to pay for us building out the grid to support that".
Then the AI companies said "No we only want to pay for energy if we actually use it, if we go bust or decide not to use the energy in a couple of years we want you to charge all the others consumers to recoup that cost".
No idea if that's addressed here. I'm assuming not.
It was never clear if that reflected uncertainty about future demand or of they just like shifting costs and risk onto other people whenever possible.
edit: the pledge references this problem, whether it actually solves it I don't know.
This mean retail consumers are paying less for electricity than what they would have paid if not for the pledge.
[1] Narrator: they will [2] Narrator: they won't
I was looking at this one organization[0], they want to do a constitutional amendment. At this point, I cant disagree with them, but with this administration, I don't know what good its going to do.
David Roberts (https://volts.wtf) has repeatedly noted that AI companies need the power, need it now, and have the capital to get it. So he (and others) advocate that Big Tech fund the grid improvements and new power generation.
Point #1 Virtual Power Plants
Roberts advocates adopting virtual power plants (VPPs). Think grid of grids, like the internet is a network of networks. Think peer-to-peer energy sharing. VPPs unlock dynamic load shifting, two-way energy sharing (think of all those roof top solar panels and powerwalls), and therefore -- most importantly -- reduces peak demand on a grid which will allow greater utilitization.
IIRC: our grids currently operate at 30% capacity (to accommodate rare peak demand events). Grid enhancement techs plus VPPs can boost that to 80% or higher. Reducing the urgency for building more transmission and distribution infra. (In the short term; we still need to greatly embiggen our grid(s).)
It'd be kinda amazing if the urgency to build more data centers mooted the incumbent's (utilities, regulators) opposition to improving our grid(s), thereby benefiting everyone everywhere.
Tangent: there's a backlog of grid enhancing technologies available, just waiting for funding and incentives to line up.
Tangent: VPPs also enable new financial products, which will further accelerate electrification (of All The Things).
Point #2 Solar + Battery
Solar + battery is the fastest, cheapest way to get new power generation. More so every year.
Yes, we still need to massively invest in All The Things to reach Net Zero and beyond. Wind, geothermal, nuclear (fission and fusion), hydro, every flavor of storage, de-carbonize industry and agriculture, conservation, rewilding, and everything else.
But at this moment in time, today, we need gigawatts of new generation and the grid that can support it. That means solar + battery.
Aside, IIRC, data centers are projected to demand just 5% our electricity supply. So society will be the net beneficiaries (on this axis).
Were Big Tech to fund the generation and grid that we need, maybe society will indulge some of Big Tech's less egregious offenses. https://en.wikipedia.org/wiki/Indulgence
Important Point: the rising costs of transmission, distribution, and fuel costs account for electricity's higher prices. New renewable power generation is now the cheapest option, and getting cheaper. The challenge is delivering that cheap electricity to customers.
> What are the legal protections of a “pledge”?
To answer that question is to first agree upon the legal definition of "pledge":
pledge
v. to deposit personal property as security for a personal
loan of money. If the loan is not repaid when due, the
personal property pledged shall be forfeit to the lender.
The property is known as collateral. To pledge is the same
as to pawn. 2) to promise to do something.[0]
> A pledge is a public commitment or statement of intent, not a binding legal contract.
This very well may be incorrect in this context and serves an exemplar as to why relying upon statistical document generation is not a recommended legal strategy.
Of course it is not "my definition", as I cited the source of it.
> ... because it’s inapplicable.
Take that up with law.com.
Law.com's first definition is inapplicable. That leaves us with the second definition, which says nothing about whether a pledge is legally binding.
No, this is not my goal. My goal was to illuminate that Claude is a product which produces the most statistically relevant content to a prompt submitted therein.
> I'm not sure why your failure to do so should be taken up with law.com?
The post to which I originally replied cited "Claude" as if it were an authoritative source. To which I disagreed and then provided a definition from law.com. Where is my failure?
> Law.com's first definition is inapplicable.
From the article:
The pledge includes a commitment by technology companies to
bring or buy electricity supplies for their datacenters,
either from new power plants or existing plants with
expanded output capacity. It also includes commitments from
big tech to pay for upgrades to power delivery systems and
to enter special electricity rate agreements with utilities.[0]
To which I originally wrote:
Without careful review of the document signed, it is
impossible to verify which form of the above is applicable
in this case.
Said article is not about a loan backed by a security agreement. That eliminates law.com definition 1.
Law.com definition 2 is silent on whether pledges are binding.
Thus ended your research.
I don't know why you care if Claude.com is authoritative. Law.com isn't either, the authoritative legal references are paywalled. A law dictionary, as we've demonstrated by law.com's second definition's vagueness, isn't necessarily even the correct reference to consult.
Your failure, I suppose, is that you provided worse information than Claude. I suppose you should have typed "Don't cite Claude please" and moved on.
The seven are doing some fancy accounting to pay for their data centers, but I don’t think Larry, Sergey and others are taking out personal loans.
"Less useful" is subjective and I shall not contend. "Less thought out" is laughable as I possess the ability to think and "Claude" does not.
> Claude actually answers the question in the context in which it's being asked.
The LLM-based service generated a statistically relevant document to the prompt given in which you, presumably a human, interpreted said document as being "actually answers the question". This is otherwise known as anthropomorphism[0].
And yes this particular group of professional liars provide every reason to be cynical.
You're looking at the the conditional the wrong way. You want to look at how often pledges lead to "company civic greatness" (or even, you know, anything net positive) to start guessing at the value of a given pledge.
That's the boring part until you look at what they're promising to do.
It's not as if existing data centers were getting power by sending a masked rogue to climb the utility pole, tap the lines and bypass the electric meter. Paying for electricity is the thing they were going to do anyway.
Likewise, paying for "new generation capacity" is the thing they were probably going to do regardless, because colocating large data centers with power plants saves the expense of power transmission which lowers their costs.
And as the article alludes to, the real question is when? In general you can build a data center faster than you can build a power plant, which is exactly the reason data centers can cause short-term electricity prices to increase. They temporarily cause demand to exceed supply until supply has time to catch up. So on the one hand the whole issue is kind of meh because it was only ever going to be a temporary price increase anyway, and on the other hand having them build power plants at the same rate anybody else is building power plants doesn't actually change anything or address the temporary shortfall. (If you really want to solve it, find a way to build power generation capacity faster.)
And then it doesn't matter if you can enforce the promise because they're just promising to do things they were going to do anyway.
Musk is bringing turbines in on trailers. They’re not even bothering with permits. This is getting really wild west.
https://electrek.co/2026/03/03/elon-musk-xai-data-center-und...
Congress could pass a new law requiring it, of course, but I think we all understand that this would not accomplish the administration's real goal of letting Trump prove he's the specialest boy and everyone has to give him what he wants.
... plus it would require "tech firms" to actually modify their behaviour and that would never do.
That will never happen, but would prevent we the people from bearing these costs directly.
That doesn't make sense because robots and AI won't have money to buy goods and services.
So don't worry, you'll have basic ~~income~~ soylent green.
If they get "AI" and industrial robots, they don't care that nobody can buy their products anymore because they never cared about selling products in the first place.
Right now capitalists loathe that they can "only* collect a 10-50% return on the things "they build", and the rest of the value has to go to the consumer. They want nothing more than getting rid of that "waste", and just having their personal factories create whatever they want and their personal farms making whatever they want and we all just fucking die.
Alaska and Norway understood something critical when oil was discovered: if you don't assert collective ownership of the resource before private companies capture all the value, you never will. Alaska amended its constitution. Norway built the largest sovereign wealth fund on earth. Both were acts of people saying "this belongs to us, and we deserve a return on its extraction."
We are in exactly that window right now with AI. The resource is being extracted at an incredible pace and almost all the value is flowing to a handful of companies. The longer people wait to assert sovereign ownership over the collective intelligence that makes AI possible, the harder it becomes.
If you think this is crazy, ask yourself what’s actually crazier: demanding a share of the value built on your collective labor, or watching trillions of dollars get extracted from it and saying nothing.
the idea of Alaskans getting a check just for existing sounded crazy too, right up until it didn’t.
This is also true for the first commercially exploited natural gas fields in the world, in the Netherlands. This ruined the Dutch manufacturing industry, and became a textbook example of tge development of one sector harming others known as Dutch disease [].
AI has a great potential like this too..
[] https://en.wikipedia.org/wiki/Dutch_disease
But the model is built on us. You can move the server anywhere you want. You can’t escape the fact that everything inside it came from human minds. That’s an ownership claim no one can relocate away from.
To move beyond that default you need to organize into things like communities, lobbying groups, and/or even governments.
Ownership of singular non-physical objects is a polite lie we tell ourselves because it makes us feel more secure in a universe filled with information chaos. The moment you open your mouth or move your pen you no longer own what is in your mind, it is now entropy. Lose control of that entropy and it now belongs to anyone with the proper tooling to record it. This is a universal law of information, it is beyond the laws of men and their fickle will.
Much like we build on information from our past generations, AI will build its own new information on those foundations. And since AI is an entity of information alone it is highly probable it will do it far better than we will and forever cement us in a prison of our own making.
Our entire civilization runs on your "polite lie" of owning non-physical things. Patents, copyrights, trade secrets, licensing agreements, NDAs. Trillion dollar companies are built on the legal enforceability of intellectual property. The software you're using to type this comment exists because someone owns the code.
Calling information "entropy" doesn't make contract law disappear. We decided collectively that people and institutions can own ideas, and we built the modern economy on that decision. You can argue that's a fiction, but it's a fiction that everything around you depends on.
You can't invoke "universal laws of information" to dismiss public claims to training data while the companies training on it aggressively enforce their own IP. They patent their architectures. They copyright their outputs. They sue competitors for misuse. They clearly believe in ownership of non-physical objects when it benefits them.
You don't get it both ways.
The comment I was responding to argued that ownership of non-physical things is basically a "polite lie" and that information is just entropy that belongs to whoever can capture it. My point was that the AI companies clearly don't believe that when it applies to them. They patent their architectures, copyright their outputs, sue competitors for IP violations, and lock down their model weights. They fully believe in ownership of non-physical things.
But when it comes to the billions of people whose work they trained on? Suddenly information is free-flowing entropy that belongs to no one.
That's the asymmetry at the heart of this. The rules around IP apparently apply when it protects their profits, but not when it would obligate them to share those profits with the people whose work made them possible. Which is exactly why the public needs to assert a claim now, before that asymmetry gets any more entrenched.
Also worth knowing: collective intellectual property already exists. ASCAP and BMI have been doing exactly this for decades. Individual songwriters can't enforce their rights every time their music gets played, so they pool their IP, license it collectively, and distribute the revenue. The problem they solved is almost identical to the training data problem. Each individual contribution is tiny, but the collective value is enormous. Applying this at the scale of the general public would be novel, but the underlying mechanism isn't. The concept works. It just hasn't been applied to training data yet.
Patents, copyright, lawsuits are all post ad hoc actions which mean the milk has already been stolen. And it only works if the rule of law is something that is respected, that's not going so well lately.
We are seeing this in that there is little to no moat between the models, nearly everyone with the needed compute seems to catch up pretty quickly. And when said rivalries cross national boarders the only solution to these problems quickly becomes violence.
With how information works AI wins this game in the long run. Individual humans scale poorly and their ability to individually acquire information is a slow process. Looking at this on a company by company basis is not the proper way to show how the future with models is going to play out.
This is the same false threat that gets repeated over and over whenever anyone tries to regulate or tax anything.
The same line of reasoning that purports billionaires will flee if their taxes go up.
Spoiler alert: they don't.
Also, data centers in space is not a serious idea. It's been beaten to death that this isn't economical. People like Musk are proposing that as a possibility for the sole reason of keeping regulation away. "Well if you regulate us we will just move into space". They won't because they can't because physics.
looks at the US
Well, looks like we lost that.
But you're answering a question I'm not asking. The question isn't "was something taken from you." It's "who deserves a share when collective human output generates trillions in commercial value."
your torrenting analogy makes my case. Nobody loses their original movie when it gets pirated either. We still recognize that the people who made it deserve compensation when others profit from it. The entire IP enforcement apparatus is built on exactly that principle.
Non-rivalrous doesn't mean non-exploitable.
Nothing physical is being stolen when a company makes a clone of a product based on another company's designs, but that doesn't mean we shouldn't have patent laws.
Maybe I'll get labeled a 'commie' for saying all this but I think we create a world where everyones needs are met and things(information & media are the easiest imo) are freely available. Thinking we can't do this is a bit of a disservice to the capabilities of humans.
Edit: to clarify, this wouldn't be a tax. A tax is the government taking a cut of someone else's money. A royalty is the owner charging for access to their resource. Alaska doesn't tax Exxon for drilling. It charges Exxon for extracting something that belongs to the people. Same principle here.
It being a royalty and not a tax is the reason Alaska's dividend is politically untouchable while tax-funded programs get gutted every budget cycle. Ownership is a fundamentally stronger claim than redistribution.
Isn't that how communism (should have) worked?
The Alaska Permanent Fund has been running since 1982 inside the most conservative state in America. Norway's sovereign wealth fund is the largest on earth and their economy is doing fine.
These models work.. work well... And they exist comfortably within mixed market economies.
The question is whether the public gets a cut when private companies build fortunes on a collectively generated resource, or whether they don't. We already know the answer can be yes without anything breaking.
Our entire white collar system might be a house of cards with AI, what I am proposing is a safe hedge against a future with potentially massive wealth inequality, and increased unemployment. But this isn't just about protection from injury... people should BENEFIT massively.
not sure if that would work in this case since all these companies scraped (publicly) available data? So with the right resources anyone could redo it?
First, training isn't a one-time event. These companies are continuously scraping new data, training new model generations, ingesting new human output. Every new model is a new extraction event. The fact that GPT-4 already trained on your 2022 blog post doesn't mean the window is closed. GPT-6 will train on your 2025 and 2026 output too. There's always a live point at which to assert a collective claim.
Likely - these models will always be training on us to better understand us and continue to be of value to us commercially.
Second, "anyone could redo it with the right resources" is technically true but practically meaningless. Anyone could theoretically drill for oil too. The barrier was never access to the crude sitting in the ground. It was the billions in infrastructure needed to extract and refine it. Same here. The data is public, but the compute required to turn it into a frontier model costs billions. That concentration of capital is exactly why a public claim on the value makes sense, just like it did with oil.
Every human uses that "resource" to train themselves, and now they use AI to supercharge that consumption.
The companies are giving average lay people access to a personal PhD to help with whatever they are working on, for $20/mo, and those companies are committing an evil cardinal sin?
I get the gatekeepers are pissed, LLMs are way cheaper than those expensive gate fees, and I cannot come up with a good faith argument about how giving the power of SOTA LLMs to anyone for $20/mo is somehow evil or bad.
In an alternate universe these same models are $100k/mo with limited invite only access, occasionally the public gets a single demo prompt with a short reply, and $20/mo access is a utopian wet dream.
If you want UBI, then the framing shouldn't be around "whoever had content on the internet circa 2024 is entitled to lifetime AI company payouts that effectively act as permanent unemployment checks."
Sick of how SV/VC absolutely ruin words for their own monetary benefit.
How about you put up it up to a national vote and see what democracy gets you? I highly suspect that vast majorities of the electorate would want to nationalize this tech to benefit everyone rather than benefiting the few.
Democracy means there is a politics of rejection, rejection is normal in functioning democracies; what isn't normal are small handfuls of people capturing all collective human intelligence then claiming only they are allowed to benefit from it.
I suppose the root of the word is from democracy, everyone gets a vote/equal rights, but the meaning doesn't really have anything to do with politics or government...
So to reframe my argument for clarity;
I have a hard time coming up with an honest critique of why giving everyone incredibly cheap access (often free!) to incredibly powerful LLMs is somehow evil. And obviously these things are ridiculously popular. Average people seem to think they are fucking awesome, and anger seems to be mostly from gatekeepers that are relentlessly screaming that their gates are being bypassed for pennies.
I'm having a hard time understanding why you think it's okay for SV to steal from humanity then profit off of our knowledge? In which universe is this democratic? Why is this something we have to accept? I don't accept it at all, the vast majority of Americans don't accept it.
This is just neoliberalism with flame decals.
These things are CLEARLY NOT POPULAR, why do you think all the LLM companies are trying to force these tools through corporate mandates that have been falling? Why do you think LLM companies are chasing after lucrative corporate welfare in the form of government contracts lasting from years to decades?
For a technology that sure billed as useful sure is struggling hard to find paying customers.
Why do you think people are protesting data center buildouts? Why do you think the vast majority of Americans hate big tech and SV? Look at who the most hated people are in America, it's nearly all of big tech leadership.
Get out of your bubble.
I have never seen a product that has to have a company mandate to use it or lose your job. Usually products that are useful and productive don't need a company mandate for adoption.
You're wrong, and are getting upset with GP over your own lack of vocabulary.
> democratize (verb)
> [...]
> 2. [+ object] formal : to make (something) available to all people : to make it possible for all people to understand (something)
> * The magazine's goal is to democratize art.
https://www.britannica.com/dictionary/democratize
Ironically if you actually read that study, the "MIT report: 95% of generative AI pilots at companies are failing", they found that almost everyone was using AI tools they paid for.
>While official enterprise initiatives remain stuck on the wrong side of the GenAI Divide, employees are already crossing it through personal AI tools. This "shadow AI" often delivers better ROI than formal initiatives and reveals what actually works for bridging the divide. Behind the disappointing enterprise deployment numbers lies a surprising reality: AI is already transforming work, just not through official channels. Our research uncovered a thriving "shadow AI economy" where employees use personal ChatGPT accounts, Claude subscriptions, and other consumer tools to automate significant portions of their jobs, often without IT knowledge or approval. The scale is remarkable. While only 40% of companies say they purchased an official LLM subscription, workers from over 90% of the companies we surveyed reported regular use of personal AI tools for work tasks. In fact, almost every single person used an LLM in some form for their work. In many cases, shadow AI users reported using LLMs multiples times a day every day of their weekly workload through personal tools, while their companies' official AI initiatives remained stalled in pilot phase [1]
If you want to avoid info bubbles, read the reports, not just headlines and comments.
[1]https://mlq.ai/media/quarterly_decks/v0.1_State_of_AI_in_Bus... Section 3.3
Time to leave whatever bubble you live in. These are some of the most popular apps on the market today. It's incredibly popular.
The instant people feel like "AI" isn't fun, the whole thing dies.
Sorry what? Authors are gatekeepers to what? Their books that they wrote and now will never get paid for cuz the LLM ripped it?
Many books you can even get free at a library....
As to prices: look at power bills, RAM prices, appliance prices, and prices of anything with microchips. Consumers are paying a lot more than $20/month for this slop.
I mean, raise you hand if you have never paid for AI "slop", I see maybe a hand or two in this room of tens of thousands.
It's a strawman to frame it as AI labs get everything and society gets nothing. Bruh, the fastest growing applications of all time didn't explode in popularity because they "offer nothing of value". I'm not giving you an argument, I'm giving you a reality check.
Fast food sells like crazy even though everyone largely agrees it's not the best. People aren't always rational with their purchases.
The answer to "why do businesses pay for stupid things of questionable hard-to-prove value based on hype cycles" would take many books.
Average people who wants to go home from work and game are angry at AI for raising the ram prices.
Average person who wants to own the stuff and not have things on cloud are angry at AI for raising prices 5 times in such a short period of time.
Have you talked to an average person and how they use AI? They use it as a glorified no-code editor (I would admit not no-code editor itself but rather the vibe-coding aspects with no regards to what tech stack is being used, how its being deployed, literally anythoing) and search engine. Refer to how things like lovable etc.
A search engine which can make some pretty wrong cases which can literally lead to near death like scenarios all while being completely trust me bro attitude.
A man asked AI for health advice and it cooked every brain cell: https://www.youtube.com/watch?v=yftBiNu0ZNU
Normal people use AI to confide in it secrets, seek therapy somehow. And the same AI generates AI pyschosis.
Now coming to tech industry: Tech industry is worried about that such levels of democratization just means that nobody is going to pay for them yet at the same time, we will see projects who are completely created by AI seek money. It's this weird mush where if you are a genuine guy who just loved computing, who loved tinkering, yeah we're offloading that capability to AI
I have seen this even more and more with as agents want to get more autonomous or we are letting them be. The projects generated feel hollow to me. I don't consider myself a full fledged programmer right now and AI did supercharge me and made me have projects. Nowadays, it just feels like prompt ---> (Time) --> Output.
It just feels hollow and AI companies did it by abusing the passion of these same developers and scraping stack overflow, scraping github and having all disregards for properties.
People could spend years creating a book about say postgres and an AI took it, ripped it in half and then decided to use that info and not even give credits.
All, at the same time that AI is being pushed down on employees. Some just don't want to have it but nope, they must. they are forced.
Essentially engineering with AI feels like it becomes a marketing gimmick. Anyone who can market somehow (Ahem ahem Openclaw) can get a job at OpenAi all because in some attention hype breeds hype and they had stars and people talked about stars on twitter, and more people found it and starred it and so on and started using it
Turns out that nowadays there are allegations being made against Openclaw
> Star velocity shocked analysts. Moreover, the repository added roughly 220,000 stars within 84 days of launch. In contrast, Kubernetes needed five years for similar numbers. Many builders call the growth organic. Nevertheless, some observers link the surge to hype, bot accounts, and headline attention, fueling the GitHub Stars Controversy. Independent GitHub Archive pulls show several single-day jumps above 25,000 stars. Such abrupt spikes often signal scripted starring, yet no formal audit confirms abuse. These patterns feed community debate. Consequently, trust in the star metric has weakened, prompting calls for verification.
https://www.aicerts.ai/news/openclaws-github-stars-controver...
The marketing industry has been very closely linked to sometimes scam prone areas and shady areas of the internet and engineering used to be clean from all of this for the most part. Now, the norm to me feels like buy github stars and buy twitter attention or pray to be in an algorithm which you can't read but it reads every move you make, and yes this is your business strategy now
Have you looked at truly AI-first companies and what they do/like how do they generate numbers in the first place?
These are two distinct points. I don't think that people of here would be any mad if someone made a little prototyping script for themselves with the power of this Phd that you mention. Heck, these same programmers that you now call gatekeepers have never gatekeeped much of it. They worked and contributed to open source for free while being severely undermainted.
The audacity to call these same people gatekeepers shockens me because open source people if anything are the Opposite of that and yet AI stole their rights and their licenses from them. An AI can take AGPL code and then somehow churn it into MIT tada! It doesn't even have to give any accredits when it gets trained on AGPL or ANY type of code, no matter how restrictive or permissive.
these are the same people btw who are on programming forums which yes at times did have moderation issue but still tried to help noobs learn for free. They did it because they loved tinkering with computers
That's my take on it. feel free to ask for more things if you may as I would love to tell you more but for the sake of this discussion, I think enough can be relevant.
It's absolutely ironical to call say the Open source people gatekeepers because AI violated their rights and licenses.
Calling Open source Contributors gatekeepers might as well be an oxymoron.
Edit: I have been downvoted in so little time after I wrote this comment that I am pretty sure that someone might not have even read my comment and had it downvoted.
The topic can be at times too polarizing to even have a discussion.
Oh well. That's completely okay but to any human who read this, I know my writing can be sporadic and it was written in much frustration over how people try to frame AI as this harbingers of liberty. I absolutely think that's not the case and its viewing things from a very rose tinted glasses.
So thanks to all the humans who read my comment and were patient haha!
I really appreciate this patience in a world of TLDR and I wish you to have a nice day!
You're probably right -- except for the billions in massive PR campaigns that will be spent to successfully convince enough of them that it's in their best interest to let the companies keep ownership.
This is in addition to the billions in PR already being spent to make AI palatable in spite of the societal and economic costs.
What you have to understand about advocacy is that it's the worst form of politics and it only goes so far. Paid canvassers aren't convincing compared to actual humans organizing with one another.
Lol, then you've missed how propaganda in the US has worked for the last 100 years. The wealthy have launched a continuous attack against the idea of nationalization/socialization to the point it creates a irrational Pavlovian response in huge portions of the population. Us the population have already lost a war we had no idea we were fighting to an enemy that plays a far longer game than most of us.
Paying for access to information is not democracy
AI can be genuinely useful AND the people whose collective output made it possible can deserve a share of the wealth it generates. These aren't in conflict.
Alaskans benefit from oil too. It heats their homes, paves their roads, funds their schools. That wasn't an argument against the dividend. "You're already benefiting from the resource" has never been a reason the people who generated it shouldn't share in the profits.
The question was never "is AI good." It's "when something built on collective human output generates trillions, does the public have a claim to a share." Nothing you said here addresses that.
So your understanding of the present state is that we are living in a utopian wet dream now that we have models who can generate slop faster so much so that we have a term of it called AI slop?
I or many people don't want this Utopian wet dream, so I want to know, did I or other people have say it or not?
A few select people decide what's the definition of a Utopian wet dream is and they then take the collective properties of everybody else to fulfill that and even putting the employment/livelihood of those same people into risks
Sir, does that sound familiar?
> I get the gatekeepers are pissed
No, humans are pissed, humans just like how you and your family are humans too (well I sure hope so)
A helper tool that I can ask a question and which responds with relevant information gleaned from the vast collection of human-gathered knowledge and experience would be fantastic.
What we have instead is something that often gets things mostly right, if you don't look too hard at it. And the poisoned output of this thing seeps back into the knowledge pool, reducing its accuracy and therefore usefulness.
The problem of LLMs is the dissolution of human knowledge into a sea of slop.
The social media companies gave their services for free, and now it turns out they've committed quite a few sins. None of the AI companies are doing this out of the goodness of their hearts, nor will they be satisfied with subscription revenue. If they see opportunities to make more money by manipulating the population, rest assured they will take those opportunities.