One of the most shocking aspects of this era of history is the number of people who not only end up accidentally resembling or aligning with the bad guys of our satire and dystopian fiction, but how many of them seem to be actively and intentionally pursuing that path. It's the Torment Nexus all the way down.
That's because there are no consequences for bad behavior, only reward. Game theory dictates that if bad behavior is a winning strategy it will be adopted and propagate until it is the dominant strategy.
The only way it stops becoming a winning strategy is if we provide consequences, but that requires taking personal responsibility for the state of the world, which was a core American value, but doesn't seem to be anymore.
How do you reconcile the belief that personal responsibility is the solution with game theoretical analysis? It seems contradictory to me.
In order to change the game theoretic outcomes, we‘d need a systemic change that affects the rewards, not a personal attitude change that will become a losing strategy in the game.
Also, do you remember how tobacco companies were invited to the table to discuss whether smoking is bad for you? Were those the days of personal responsibility or was it even before that?
I will dare say, it's a question of what happens to Mario's brother. Jury nullification in that is the best message that could be sent to the populace.
I'm pretty sure giving everyone the belief that they can murder with impunity as long as the victim is undesirable enough would be the absolute worst thing that could happen right now.
One of the dystopian traits I was hinting at in my original comment is the acceptance of mass murder as long as it occurs in the boardroom and nets an extra few cents for the shareholders. I won't defend or justify what "Mario's brother" supposedly did, but he has inflicted much less pain and death on the world than the man he is accused of shooting and I don't really think there is any room to debate that.
Hypothetically if laws lose their legitimacy because nobody is willing to enforce them (at least some people have that perception) and the political system is designed in such a way as to make any meaningful change near impossible what is there left to do?
And more generally the fundamental reason we agree to operate under laws is because these laws are expected to improve society as a whole. But if those laws instead start enabling and protecting bad behavior then they're doing the exact opposite.
It's fairly obvious that much of what the more sociopathic corporations do today will be illegal in the future, but changes in social opinions tend to predate changes in the law by quite some time. For the obvious extreme there - slavery was completely legal. Society began to believe that such a thing was no longer fit for society, and consequently acting against it, long before it was outlawed.
The fact that such behavior might become acceptable (at least amongst a significant section of society) indicates a systemic failure in the socio-economic system. IMHO its more a of symptom. Like labor and anarchist related political violence back in the first Gilded Age back in the late 1800s.
Massively increasing inequality and giving too much political power to corporate robber barons has its costs. If nobody is willing to keep them in check the appearance of some sort of “vigilantism” seems hard to avoid. Not implying that its a good thing or that political violence really ever led to positive change historically..
We can already murder with impunity. Just get a few people together, form a company, and you can do whatever. Even crimes against humanity won’t get you in prison, probably. At worst you’ll lose some cash and be demoted to “normal person”.
How many people has UHC killed? I don’t know, it’s really hard to measure. Besides the people killed because they didn’t receive funding for care, there’s also the plethora of practices insurers enforce. Some, maybe most, of those practices are non-optimal, so some subset of people are dying that shouldn’t. Oh well.
That belief is already commonplace, and has been vigorously tested among the sex worker, queer, Black and Native communities and proven correct. I don't see why we should be any more concerned about adding "rich white men" to the pile than we are about any of the other disposable demographics in our society.
The fact that rich engineers on hacker news would be flirting with bolshevism as their ideology is just endlessly funny to me. I know it’s just people parroting the emotional “vibe” of their political tribe on any given day, but it’s so ironic.
Beyond the obvious moral decay of cheering on murder at all, and the fact you’re in the privileged class of the richest nation on earth, the idea of targeting the replaceable middle managers of said system is so silly. As if committing random acts of terrorism will somehow force Americans to democratically design a better system? Fear is just another recipe for more ballooning costs (see the TSA).
I guess I find this so amusing because leftists love to fetishize European healthcare without understanding in European countries the government is much more aggressive about denying care than any US insurer. They actually have to keep costs sane for their system to continue existing.
> They actually have to keep costs sane for their system to continue existing.
That is also the case for US insurers. The only difference is if the government denies life saving treatments, people protest. If private insurers do so, people have no recourse.
In both situations you have zero recourse. In fact the US Government is less responsive to protest than US businesses are.
US healthcare is one of the most complicated systems of adverse incentives and tangled byzantine public/private spiderwebs ever created. To kill random people involved at 15 layers of abstraction away from the actual root causes thinking that will somehow make it better is probably the dumbest idea I've ever heard.
The only way out is to hold executives personally responsible for the actions of their companies, and politicians for the results of their policy.
Sam Altman should receive the same treatment as Aaron Swartz. Actually, he should be punished much more severely since the scope of his copyright infringement makes Aaron's seem like child's play.
"Smaht"[1] people learn to game the system and scam others for momentary benefit.
The worse side that is that we're all guilty of that system, to some degree, even if only by enabling it.
I'm also 100% sure that this is what drives civilizations to the ground.
1. Smaht is a term I use to describe people who think they're smart but they're actually extremely stupid. A lot of smaht people have degrees and diplomas which further fuels their delusion of intelligence.
Sci-Fi Author: Inspired by human atrocities, I present to you my new novel: God Emperor of Dune.
Tech adjacent blogger: Hey guys here me out I love that we're building "starships" but it would also be spiffy if we end democracy and appoint a God Emperor!
I’ve had that same thought. It’s wild how often real-world decisions echo the exact warnings from sci-fi and satire. Sometimes it feels like people read dystopian fiction not as a cautionary tale, but as a roadmap. The "Torment Nexus" joke stopped being funny a while ago because it keeps getting closer to reality.
I really wish I could know if they are earnestly cosplaying Lex Luther or if they are just deluded. Of course a good Lex Luther cosplay would involve misdirection so it’s basically impossible to know. It doesn’t really matter which one it is because the outcome is similar but it would be very gratifying to know.
As opposed to those "unaligned" communist open-source models. As a proud freedom-loving citizen of the West you wouldn't want to support those now would you?
100%, though I still feel as though open training data will eventually become a thing. It'll have to be mostly new data, synthetic data, or meticulously curated from public domain / open data.
Synthetic training data sets, even robotically-acquired real world "synthetic" data, can rapidly create training sets. It's just a matter of coordinating these efforts and building high quality data.
I've made a few data sets using Unreal Engine, and I've been wanting to put various objects on turn tables and go out on backpack 3D scan adventures.
I just finished reading "Careless People" and the tone is shockingly similar to the one Zuckerberg loved to use. It reminds me of that Silicon Valley scene where every startup wants to "make the world a better place".
As someone who is both expected to keep creating information to train AI while being stripped from the fruit of my labour by it, I find it sickening.
Comes with a free US government backdoor to all of the foreign citizens' data and AI usage.
Though of course this is already the status quo for all US companies abroad, so you have to give props to OpenAI for spelling it out explicitly: Give up what remains of your digital sovereignty to the US government and you get a small piece of the AGI pie.
The pattern for basically every small nation is "choose of which superpower you wish to be a client." From that patron you get some level of benefit. Not aligning with any either doesn't work (you get attacked) or means you get no benefit (and eventually get pushed into obscurity and instability.)
You can make a lot of complaints about America but we have, looking back on history, been nicer than any other patron. Other good evidence includes the fact that europe is still standing (paying to rebuild) and her extravagant welfare states of the past decades, subsidized largely by American defense spending.
> The pattern for basically every small nation is "choose of which superpower you wish to be a client."
This is straight up Russian mentality.
> extravagant welfare states of the past decades, subsidized largely by American defense spending
This sounds to me like a US partisan narrative rather than anything else. It’s a nice story, because it strokes the American ego, but I’ve yet seen it backed up by serious analysis. Most likely it’s just a story.
I agree with most of what you said. America has been a great ally, mostly by allowing her allies to flourish independently of herself. The US did whatever she wanted to do, and so did her allies. This was a great benefit to all involved.
> subsidized largely by American defense spending.
This part is in my opinion ahistoric. US wars have not been popular in Europe. We did not want a war in Afghanistan or Iraq, we supported an ally calling for defense from terror. American war machine spending is rooted in her own desire for hard power, not pleas from her allies.
All of this is coming to an end. Not because the US is retracting. I think most of the west would accept a more nationally interested US, but because the US is starting to see her allies as vassals that she should control. She is realigning as a traditional power, like the USSR.
We are not vassals, we are independent nations seeking our own happiness.
How can a glorified NLP app be equated with being the backbone of economic development and a path to AGI ? So many people have been fooled by marketing.
Honestly though, we have a much bigger issue with climate change in the medium to long run and it doesn't really matter what our governments and companies do with stats and spyware. If anyone thinks we can stop and deal with the climate when it becomes a bigger problem, just take a look at our track record so far.
(only mentioning climate change to offer perspective)
> How can a glorified NLP app be equated with being the backbone of economic development and a path to AGI ? So many people have been fooled by marketing.
Regulators are still figuring out this “AI” and oAI must move into as many market to sustain their valuation and future before regulations start to close many open doors.
Also, when entire EU comission makes “AI” a core focus, all other governments are having a FOMO, which is the most fertile opportunity to entrench oneself quickly before everyone realises the smoke and mirror of “productivity gain” song means just making another layer of middleman mandatory for everything(see Apple pushing towards modifying Safari to be AI first).
Also what climate change? Everyone was being shamed into indignation recently for their carbon footprints, only to wake up to massive power infra expansion and Nvidia/Amazon/Msft announcing that everything is on the table including burning more fossil fuel to power the energy demand(utilities are usually often govt controlled and hence a social cost overall).
The climate change that, if left unchecked, will almost certainly lead to the death of much of humanity, and the majority of life on Earth. Hundreds of millions of climate refugees knocking at your door.
> The Trump administration plans to rescind Biden-era AI chip curbs as part of a broader effort to revise semiconductor trade restrictions that have drawn strong opposition from major tech companies and foreign governments, according to people familiar with the matter.
This doesn't mean anything. He won't get enough likes on Xitter tomorrow and will flip-flop to 1000% tariffs or whatever else comes to his senile mind.
This unstable circus of a government can't be trusted.
On several of Tyler Cowen's recent podcasts, he has said essentially "there are really only two countries that have AI, China and the USA. Does this mean that other countries (like Peru) won't really have a functioning, powerful government when AI runs everything".
> As you know, not many countries have serious AI companies, and even those in Europe may or may not last. They’re not obviously mega profitable. Let’s say you’re the government of Peru, and you can turn over your education system to some foreign, maybe American, AIs. You can turn over how your treasury is managed to the AIs. You can turn over your national defense to the AIs. None of these are Peruvian companies most likely. In the final analysis, are we even left with the government of Peru? Or has it, in some sense, been pseudo privatized to the companies that are running the structures, and indeed to the AI itself?
Interesting to have OpenAI offer up AI infra so other countries are not at quite as large a disadvantage. Also really good for their business.
IMO that analysis is shortsighted when looking at other technologies. Peru’s government would grind to a halt with say, windows/osx, excel, chrome, email etc. They are all tools that enable work. I don’t see AI being categorically different.
In this hypothetical world where AI runs the treasury, is the US now in a massively better position to make treasury related decisions? Maybe? Does the US gov have a remote chance of abiding by these decisions? Etc.
I can see Peru being disadvantaged if they don’t use AI, but if they contract out and set up their own stuff that they didn’t actually build - how’s that really worse? I feel like they let the US spend hundreds of billions in development costs and can now reap the rewards.
> They are all tools that enable work. I don’t see AI being categorically different.
You don’t see the difference between a self-contained product, and a foreign subscription service with no influence over what it is delivering and the privacy and data sovereignty implications? Let alone the vast array of subtle manipulation possibilities in responses?
By and large, those technologies do not come with an always-on umbilical that leads out of the borders of those countries. It is relatively easy to build out capacity, unlike with AI that requires extremely specialised hardware in vast quantities.
You can't be seriously considering fancy autocomplete word guessers are replacing governments when Musk can't even get Grok to stop telling Twitter users what a moron he is.
Anthropic is allowing the US government to use their services. This includes various intelligence organisations and their data analysis they presumably use to target strikes.
> fancy autocomplete word guessers are replacing governments
UK has had them in government since 2022, or maybe since Brexit/ Teresa May with her nickname Maybot.
The decline in quality of governance has been so severe, that I’d wager you would not see a difference. Both sides of the isle seem to be full of unintelligent or inexperienced people that do not believe in anything or have a vision
> You can turn over your national defense to the AIs. None of these are Peruvian companies most likely. In the final analysis, are we even left with the government of Peru?
Folks, this has already been happening for decades, western consultancies and think tanks have been pushing for privatisation and outsourcing to American firms and as a result many governments, like UK, have been hollowed. In many cases they haven’t got a grip and the country is running on autopilot.
As the consultancies replace employees with AI, the outcome you talk about will be achieved, in about 5 years. No far fetched future required
Heading there? Facebook has been a kingmaker for a decade. Musk runs DOGE. Most big companies can bully smaller administrations when they feel the need.
I dont get the proposition, they want to build DCs in partnering countries to run GPT on? Who is this useful for, except for OpenAI to get lower latency connections to their customers?
Not latencies, think data privacy / keeping queries and data from leaving sovereign borders. This way, if there is some local instance / everything is local than the datacenter and service are subject to local laws and regulations (and alternatively you're not subject to foriegn the laws and regulations (and agencies).
That's not quite correct. The "sovereignty" pitch here is largely illusory when dealing with a US-based company like OpenAI.
The US CLOUD Act (Clarifying Lawful Overseas Use of Data Act) explicitly gives US authorities the power to compel US-based companies to provide data stored on servers, regardless of where those servers are physically located. This effectively undermines any meaningful data sovereignty claims.
Consider the actual arrangement being proposed:
- OpenAI (US company) maintains control of the infrastructure
- OpenAI controls the models and their development
- OpenAI maintains the security protocols and access rights
- The data merely sits physically within national borders
This isn't sovereignty - it's a limited hosting arrangement that remains fully under US legal jurisdiction. US intelligence agencies can still access this data through legal mechanisms that bypass the host country's laws entirely.
locality is good for resilience and latency but for privacy? how does it work?
How can one audit that the bytes going from a DC in country A to a DC in the US is not the user queries but some telemetry data for example? Presumably you don't get to look at the unencrypted packets
See, you just don’t get it, we will only be free when we get rid of politicians and have referendums on all legislation.
But you can sell options on your family’s votes.
Once someone sells a vote, they are in vote debt, and can default by voting a different person than they agreed.
So now you have to have a credit rating, but for voters. Then you need to have Voter Default swaps, which can be Bundled into Voter Default Obligations, Of VDO’s. And then you can have Synthetic Voter Default Swaps and ahead of a major election you can do a Big Short.
The Stargate link is notable since that has received a large amount of backing from the United States government, who hasn’t been friendly with other countries lately.
Stargate has no US government funding. It was latched onto by Trump to pretend he was immediately making some “deals”. But the whole thing is an illusion of pre-existing projects and investments that pre-date the last election.
So let me get this straight: countries fund the infrastructure, shoulder the risk, dole out taxpayer money to the for-profit arm of OpenAI, weaken privacy laws, and hand over taxpayer data for…nothing? It just reads like a “hey gullible suckers, give us your land/money/data and we’ll let you slap our logo on stuff until it’s no longer economically convenient for us to do so, at which point we’ll demand you subsidize us because we can claim we’re indispensable/too big to fail” grift to me, unless I’m missing something.
They mention a good point (which probably most countries already realized), but the obvious answer is to invest in lowercase open AI, not uppercase OpenAI.
So that's just (or "just") locating the inference infrastructure inside the user's country? All operations, deployment, all training, tuning and development, contract negotiations remain the same?
This is a genius move to lock in revenue from countries lacking the technological infrastructure and capital to develop and run their own "safe" (for the local junta) models. Doubly so that OpenAI are experts in censorship - I mean "alignment" - and can help local authorities impose a localized censorship regime. The logical next step is going hard on promoting "AI Safety" and legislating the use of certified approved censored models in each locale, and criminalizing the use and possession of unapproved models, the same way certain JPEG files carry multi year prison sentences or how possession of certain books in certain countries carries prison time.
I’ll try, but not succeed with a view from Australia..
Companies and governments have been concerned about data and AI sovereignty, and chip (processing) access. The new risks imposed by the USA are increasing this concern / push.
So, it’s hardly surprising that Sama is getting a lot of calls for local instances.
However, if the data etc. moves back to the USA this is exactly the opposite of the control companies and governments are looking for.
Most of the commentary is presuming to know something about OpenAI’s motivations. That’s not honesty; it’s just an opinion. So my question stands. Does anyone have a positive opinion?
Here’s a take. For those of us who use their tools in our day to day, we might take for granted that we have the existing and new infrastructure to support that product. Is it more good than bad that other parts of the world could reach beyond their current grasp? I hope so. It might be.
It’s not a requirement. If it’s just the culture of HN to dunk on certain companies or products, then it is. This place doesn’t belong to me any more than it belongs to you. I am hopeful though that we could encourage more diversity of opinion here. Otherwise it’s exhausting.
And yes, I do, and it’s shared in a different comment. Search if you care to read it.
Interesting, I've had the opposite feeling of AI being super hyped throughout the industry, with tons of positivity and not a whole lot of reflection or criticism.
You provide the capital and the data, we'll co-own the data centers share the models until Trump and the US government decide to shut it off as a bargaining chip.
> Through formalized infrastructure collaborations, and in coordination with the US government, OpenAI will:
> Partner countries also would invest in expanding the global Stargate Project—and thus in continued US-led AI leadership and a global, growing network effect for democratic AI.
Yeah, good luck with that pitch... I have to assume that the target market for this page is not other countries, but the US leadership.
Hmm, the cynic in me reads this as move fast and capture market(+regulation) before new regulation is setup to thwart the likes of GDPR and other privacy acts. When something is new and regulators are having hard time understanding the consequences and future risks, it is most efficient and cheap to capture the market. Once the fallouts start, regulations strike but by then the early players are too big and well established to deal with anything, while the new and smaller players get crushed under compliance and consequences of the early big players’ shenanigans.
This is simultaneously why most people desperately want to invest in OpenAI and at the same time why all the best gen AI researchers want to work for anthropic. The less you understand the more impressive this seems. Conversley the more you understand the more embarrassing this seems.
> We want to help these countries, and in the process, spread democratic AI, which means the development, use and deployment of AI that protects and incorporates long-standing democratic principles (…) Likewise, we believe that partnering closely with the US government is the best way to advance democratic AI.
The current US government? To protect “long-standing democratic principles”? Give me a break.
> We’ve heard from many countries asking for help in building out similar AI infrastructure—that they want their own Stargates and similar projects.
Who is this for exactly? The thing about reneging on your agreements and treaties and threatening and demonizing everyone around you is that they learn not to trust you. US-led AI sounds terrible, it would never pass muster in Canada. Neither in the EU, China, India, Brazil... Like, you Cannot entrust your governments functioning on the US anymore, you can just get cut off at any point for no reason.
Brazil it's actually building a project to give tax breaks and change/simplify a few regulation for datacenter projects. Import tariffs will also be 0 for them...
Finance Minister it's in the California trying to bring investments from the big techs... He met with Jensen Huang already.
I wouldn't doubt if Brazil might be interested.
TikTok is also interested in building a datacenter in Ceará, Brazil, as part of this project.
How is this different from NGOs flooding poor countries with food aid and causing prices to crash, local farmers to go bankrupt and then become dependent on that same aid? There are 2nd and 3rd order effects on innocent partnerships everywhere.
Well I guess the time scale is what determines the degree to which the distinction becomes opposition. AI is likely to persist for tens or hundreds of thousands of years in some form. Are any of today's nation states built to last that long? I think we all know the answer.
If you have AI which is in the service of an entity which proclaims itself to be the sole franchise of government authority over a given landmass, it is strictly incorrect to say that this AI is "for the country", because it's perfectly plausible (and on sufficiently long time scales, inevitable) that the country will want to evolve, replace, or deprecate that entity.
I agree that “AI for governments” is much more accurate, just saying that diametrically opposed doesn’t really capture the relationship between the two concepts well
This is disturbing to read and wonder what other countries are going to want “democratic” AI developed in partnership with and “led by” the US and Trump.
Not sorry, I don't trust anything with Trump's name attached to it. I have to live here, but I don't have to like it, or trust him or anyone attached to him.
I'm reading this in the same voice as Helldivers 2 "managed democracy"
The only way it stops becoming a winning strategy is if we provide consequences, but that requires taking personal responsibility for the state of the world, which was a core American value, but doesn't seem to be anymore.
In order to change the game theoretic outcomes, we‘d need a systemic change that affects the rewards, not a personal attitude change that will become a losing strategy in the game.
Also, do you remember how tobacco companies were invited to the table to discuss whether smoking is bad for you? Were those the days of personal responsibility or was it even before that?
A rookie gangster shot a high-tier gangster — why is that a me problem? Gang-on-gang violence is a daily occurrence.
It's fairly obvious that much of what the more sociopathic corporations do today will be illegal in the future, but changes in social opinions tend to predate changes in the law by quite some time. For the obvious extreme there - slavery was completely legal. Society began to believe that such a thing was no longer fit for society, and consequently acting against it, long before it was outlawed.
Massively increasing inequality and giving too much political power to corporate robber barons has its costs. If nobody is willing to keep them in check the appearance of some sort of “vigilantism” seems hard to avoid. Not implying that its a good thing or that political violence really ever led to positive change historically..
How many people has UHC killed? I don’t know, it’s really hard to measure. Besides the people killed because they didn’t receive funding for care, there’s also the plethora of practices insurers enforce. Some, maybe most, of those practices are non-optimal, so some subset of people are dying that shouldn’t. Oh well.
Beyond the obvious moral decay of cheering on murder at all, and the fact you’re in the privileged class of the richest nation on earth, the idea of targeting the replaceable middle managers of said system is so silly. As if committing random acts of terrorism will somehow force Americans to democratically design a better system? Fear is just another recipe for more ballooning costs (see the TSA).
I guess I find this so amusing because leftists love to fetishize European healthcare without understanding in European countries the government is much more aggressive about denying care than any US insurer. They actually have to keep costs sane for their system to continue existing.
That is also the case for US insurers. The only difference is if the government denies life saving treatments, people protest. If private insurers do so, people have no recourse.
US healthcare is one of the most complicated systems of adverse incentives and tangled byzantine public/private spiderwebs ever created. To kill random people involved at 15 layers of abstraction away from the actual root causes thinking that will somehow make it better is probably the dumbest idea I've ever heard.
Sam Altman should receive the same treatment as Aaron Swartz. Actually, he should be punished much more severely since the scope of his copyright infringement makes Aaron's seem like child's play.
"Smaht"[1] people learn to game the system and scam others for momentary benefit.
The worse side that is that we're all guilty of that system, to some degree, even if only by enabling it.
I'm also 100% sure that this is what drives civilizations to the ground.
1. Smaht is a term I use to describe people who think they're smart but they're actually extremely stupid. A lot of smaht people have degrees and diplomas which further fuels their delusion of intelligence.
Tech Company: At long last, we have created the Torment Nexus from classic sci-fi novel Don't Create The Torment Nexus
(@AlexBlechman on twitter)
Tech adjacent blogger: Hey guys here me out I love that we're building "starships" but it would also be spiffy if we end democracy and appoint a God Emperor!
Quite often this dystopian 'fiction' is just a biography with the names and place rewritten. A scary number of people are rather anti-human.
https://chatgpt.com/share/681c31e8-67f8-8011-a4b0-2bed9d4da7...
I'm reminded of the first half of this wonderful short-story that was shared on HN a year back https://www.fortressofdoors.com/four-magic-words/
Open weights and code and models? That's the only way to ensure sovereignty.
I think this company is a walking oxymoron.
Synthetic training data sets, even robotically-acquired real world "synthetic" data, can rapidly create training sets. It's just a matter of coordinating these efforts and building high quality data.
I've made a few data sets using Unreal Engine, and I've been wanting to put various objects on turn tables and go out on backpack 3D scan adventures.
Someone will have to pay for it, though.
As someone who is both expected to keep creating information to train AI while being stripped from the fruit of my labour by it, I find it sickening.
Tech companies only care about growth. They only care about anything else insofar as it supports growth.
Well, OpenAI, I think you are mixing up your own backend for economic growth with everyone’s!
there is no data sovereignty if there's a US entity at the top
Though of course this is already the status quo for all US companies abroad, so you have to give props to OpenAI for spelling it out explicitly: Give up what remains of your digital sovereignty to the US government and you get a small piece of the AGI pie.
You can make a lot of complaints about America but we have, looking back on history, been nicer than any other patron. Other good evidence includes the fact that europe is still standing (paying to rebuild) and her extravagant welfare states of the past decades, subsidized largely by American defense spending.
This is straight up Russian mentality.
> extravagant welfare states of the past decades, subsidized largely by American defense spending
This sounds to me like a US partisan narrative rather than anything else. It’s a nice story, because it strokes the American ego, but I’ve yet seen it backed up by serious analysis. Most likely it’s just a story.
> subsidized largely by American defense spending.
This part is in my opinion ahistoric. US wars have not been popular in Europe. We did not want a war in Afghanistan or Iraq, we supported an ally calling for defense from terror. American war machine spending is rooted in her own desire for hard power, not pleas from her allies.
All of this is coming to an end. Not because the US is retracting. I think most of the west would accept a more nationally interested US, but because the US is starting to see her allies as vassals that she should control. She is realigning as a traditional power, like the USSR.
We are not vassals, we are independent nations seeking our own happiness.
Honestly though, we have a much bigger issue with climate change in the medium to long run and it doesn't really matter what our governments and companies do with stats and spyware. If anyone thinks we can stop and deal with the climate when it becomes a bigger problem, just take a look at our track record so far.
(only mentioning climate change to offer perspective)
Regulators are still figuring out this “AI” and oAI must move into as many market to sustain their valuation and future before regulations start to close many open doors.
Also, when entire EU comission makes “AI” a core focus, all other governments are having a FOMO, which is the most fertile opportunity to entrench oneself quickly before everyone realises the smoke and mirror of “productivity gain” song means just making another layer of middleman mandatory for everything(see Apple pushing towards modifying Safari to be AI first).
Also what climate change? Everyone was being shamed into indignation recently for their carbon footprints, only to wake up to massive power infra expansion and Nvidia/Amazon/Msft announcing that everything is on the table including burning more fossil fuel to power the energy demand(utilities are usually often govt controlled and hence a social cost overall).
Except USA banned export of GPUs to like half of the European Union, let alone third-world countries.
https://archive.is/2eLzj
> The Trump administration plans to rescind Biden-era AI chip curbs as part of a broader effort to revise semiconductor trade restrictions that have drawn strong opposition from major tech companies and foreign governments, according to people familiar with the matter.
This unstable circus of a government can't be trusted.
See https://conversationswithtyler.com/episodes/chris-dixon/
> As you know, not many countries have serious AI companies, and even those in Europe may or may not last. They’re not obviously mega profitable. Let’s say you’re the government of Peru, and you can turn over your education system to some foreign, maybe American, AIs. You can turn over how your treasury is managed to the AIs. You can turn over your national defense to the AIs. None of these are Peruvian companies most likely. In the final analysis, are we even left with the government of Peru? Or has it, in some sense, been pseudo privatized to the companies that are running the structures, and indeed to the AI itself?
Interesting to have OpenAI offer up AI infra so other countries are not at quite as large a disadvantage. Also really good for their business.
In this hypothetical world where AI runs the treasury, is the US now in a massively better position to make treasury related decisions? Maybe? Does the US gov have a remote chance of abiding by these decisions? Etc.
I can see Peru being disadvantaged if they don’t use AI, but if they contract out and set up their own stuff that they didn’t actually build - how’s that really worse? I feel like they let the US spend hundreds of billions in development costs and can now reap the rewards.
You don’t see the difference between a self-contained product, and a foreign subscription service with no influence over what it is delivering and the privacy and data sovereignty implications? Let alone the vast array of subtle manipulation possibilities in responses?
You can't be seriously considering fancy autocomplete word guessers are replacing governments when Musk can't even get Grok to stop telling Twitter users what a moron he is.
UK has had them in government since 2022, or maybe since Brexit/ Teresa May with her nickname Maybot.
The decline in quality of governance has been so severe, that I’d wager you would not see a difference. Both sides of the isle seem to be full of unintelligent or inexperienced people that do not believe in anything or have a vision
what an oxymoron.
this is testament how good grok is.
Folks, this has already been happening for decades, western consultancies and think tanks have been pushing for privatisation and outsourcing to American firms and as a result many governments, like UK, have been hollowed. In many cases they haven’t got a grip and the country is running on autopilot.
As the consultancies replace employees with AI, the outcome you talk about will be achieved, in about 5 years. No far fetched future required
https://us.macmillan.com/books/9780374615369/wheretheaxeisbu...
Source?
I've read so many sci-fi stories where big tech corporations have similar control over people as countries. Now we are actually heading there.
I'm both excited and a bit worried about the future.
The US CLOUD Act (Clarifying Lawful Overseas Use of Data Act) explicitly gives US authorities the power to compel US-based companies to provide data stored on servers, regardless of where those servers are physically located. This effectively undermines any meaningful data sovereignty claims.
Consider the actual arrangement being proposed:
This isn't sovereignty - it's a limited hosting arrangement that remains fully under US legal jurisdiction. US intelligence agencies can still access this data through legal mechanisms that bypass the host country's laws entirely.How can one audit that the bytes going from a DC in country A to a DC in the US is not the user queries but some telemetry data for example? Presumably you don't get to look at the unencrypted packets
Helping people do more? Scaling our ability to create and produce?
Sadly, none of these things ever made us happier as humans.
Free markets concentrate wealth and power.
Concentration of wealth and power is antithetical to democracy.
But you can sell options on your family’s votes.
Once someone sells a vote, they are in vote debt, and can default by voting a different person than they agreed.
So now you have to have a credit rating, but for voters. Then you need to have Voter Default swaps, which can be Bundled into Voter Default Obligations, Of VDO’s. And then you can have Synthetic Voter Default Swaps and ahead of a major election you can do a Big Short.
Have you considered that this proposition is even too ridiculous for current reality?
Brilliant in a Bond villain way
I'm deeply pessimistic.
lmao, is there a single soul at openai who truly believe this bullshit?
Are they so high on their own supply they can't even tell they're becoming a parody of a black mirror evil corp?
Companies and governments have been concerned about data and AI sovereignty, and chip (processing) access. The new risks imposed by the USA are increasing this concern / push.
So, it’s hardly surprising that Sama is getting a lot of calls for local instances.
However, if the data etc. moves back to the USA this is exactly the opposite of the control companies and governments are looking for.
So, fair proposal, wrong execution.
Most of the commentary is presuming to know something about OpenAI’s motivations. That’s not honesty; it’s just an opinion. So my question stands. Does anyone have a positive opinion?
Here’s a take. For those of us who use their tools in our day to day, we might take for granted that we have the existing and new infrastructure to support that product. Is it more good than bad that other parts of the world could reach beyond their current grasp? I hope so. It might be.
To increase shareholder value?
And yes, I do, and it’s shared in a different comment. Search if you care to read it.
You provide the capital and the data, we'll co-own the data centers share the models until Trump and the US government decide to shut it off as a bargaining chip.
> Partner countries also would invest in expanding the global Stargate Project—and thus in continued US-led AI leadership and a global, growing network effect for democratic AI.
Yeah, good luck with that pitch... I have to assume that the target market for this page is not other countries, but the US leadership.
The current US government? To protect “long-standing democratic principles”? Give me a break.
https://erinkissane.com/meta-in-myanmar-full-series
Who is this for exactly? The thing about reneging on your agreements and treaties and threatening and demonizing everyone around you is that they learn not to trust you. US-led AI sounds terrible, it would never pass muster in Canada. Neither in the EU, China, India, Brazil... Like, you Cannot entrust your governments functioning on the US anymore, you can just get cut off at any point for no reason.
So who's this for?
Finance Minister it's in the California trying to bring investments from the big techs... He met with Jensen Huang already.
I wouldn't doubt if Brazil might be interested.
TikTok is also interested in building a datacenter in Ceará, Brazil, as part of this project.
If you have AI which is in the service of an entity which proclaims itself to be the sole franchise of government authority over a given landmass, it is strictly incorrect to say that this AI is "for the country", because it's perfectly plausible (and on sufficiently long time scales, inevitable) that the country will want to evolve, replace, or deprecate that entity.