Alright. This is it. OpenAI has officially attracted my wrath and they’re going to get it. Usually, this blog is reserved for professional, well-crafted entries. Well, cya pretense.

Introduction

As you may have heard, OpenAI has been signing a ton of compute deals lately. The amount of money being thrown around is truly fantastic - as in unbelievable. They have made a total of $1.4T deals on compute and hardware. Everyone is calling out the obvious-as-hell bubble.

What you may not have heard is that “Sarah Friar, the CFO of OpenAI, says the company wants a federal guarantee to make it easier to finance massive investments in AI chips”, so, they’re preemptively asking the government to guarantee them bailouts. By framing the whole thing as a national security issue.

The classic move of “privatize profits, socialize losses”. They are literally going by the 2008 financial bubble winners playbook.

The Scale

Let’s examine the amount of money that’s being asked for here. $1.4T, just how much is that? Well, the USA has something like 350 million people. So this is $4000 per American citizen - and yes, this includes 2-year-old little Joanna. Every family of 4 pays them ~$16,000. This, when 40% of Americans can’t even afford a $400 emergency.

Let’s think about what else could be done with this money.

• It comes pretty close to erasing all of: student debt ($1.7T),
• easily solves all of homelessness ($20B)
• provide a year of healthcare for everyone
• eliminate child poverty ($4380 per family could lift millions out of poverty) cost: $100-$150 billion / year
• free college education ($60-$95 billion / year)

I mean, holy shit. We are talking about a lifechanging amount of money for a huge percentage of Americans here. Something that could have measurable quality of life increase for everyone.

This money is going to come from cuts in social programs, increased taxes on working people, inflation from money printing.

The Mechanism

I want to lay out the logistics for how when OpenAI’s CFO asks for federal guarantees to “make it easier to finance massive investments in AI chips,”, this translates to “getting the money from the public”.

So, there are a few interpretations:

• Loan guarantees - Government backs loans to OpenAI/partners, meaning if they default, taxpayers cover it
• Infrastructure guarantees - Government commits to buying/using the infrastructure if private deals fall through
• Implicit bailout promise - Signaling that OpenAI is “too important to fail”

There is a chance I’m missing something, but these are all the interpretations I’m seeing. Any of these effectively mean:

“We want to make massive commitments we might not be able to afford, and we want assurance we won’t face consequences.”

So, what actually happens:

1. Let’s say OpenAI makes a deal with Oracle for $200B in compute. Oracle really doesn’t expect OpenAI to be able to pay them back
2. OpenAI does use this $200B of compute.
3. OpenAI can’t make good on the payments it owes.
4. Without guarantees: Oracle eats the loss, investors lose money, OpenAI potentially bankrupts
5. With guarantees: Government steps in, for example by Direct payments to Oracle to “preserve critical AI infrastructure”, Nationalization of the infrastructure, Bailout of OpenAI itself Either way, taxpayers end up covering the bill via some combination of money printing, cutting social programs, debt (future tax burden)

And, having a guarantee in advance that costs will be paid, every compute provider is incentivized to make the largest deals possible, because there is no chance of them eating the cost. It’s an infinite money glitch for them. They can make deals as large as they want and know that in the end, the taxpayers will pay it in some way.

Framing As National Security Issue

I bet you’ve been hearing a lot how if the US loses to China, everyone is fucked forever. It’s been a powerful message to aid in accelerating development with minimal guardrails, increase public belief that this truly is a national security issue.

Meanwhile, OpenAI also likes to frame their LLMs as mere products for business use. They will up and downplay the safety issues as it conveniences them, really. But yeah, this absolutely is a national security issue. There are still dissidents who don’t spread the word that the current US government is the greatest government in human history. They certainly are the greatest at being openly and transparently corrupt. They don’t even make attempts at hiding it anymore. This technology is a surveillance states wet fucking dream. Like holy shit does the NSA cream itself every night thinking about how they can now semantically analyze every message to scan for people that say bad things about the president. Or, you know, have AI-enabled cameras all over everywhere to track everyones IRL movements at all times. They can delete in real-time any comment criticizing the government from the internet if its posted on a site hosted by a company in the US, or just have Google not search-index it. It is amazing to destroy any coordination against the US government, making sure it will reign supreme for another 1000 years.

Anyhow, if I was the Trump administration, I’d obviously see this opportunity for increasing the surveillance state. If OpenAI can’t pay its Stargate, I’d cut a deal with them to gain ownership of the thing, then have all that compute be used for national security matters (like what I laid out in the previous paragraph)

The capabailities are obviously there to use AI for many surveillance tasks even in its current state. For inspiration, they don’t really need to look further than China. There already was a recent push towards AI-enabled surveillance cameras which was heavily criticized and fought against by Louis Rossmann. So, it’s quite obviously on the table.

So yeah, this is great for “national security”. Nobody will ever be able to do anything against the state when properly implemented. This is what you’re really funding with the $4000 you’re donating to OpenAI & investors right now. I’m just sort of imagining that most US citizens are exhibitionists who like being watched jerking off by the government, otherwise they’d have to be less tolerant of this shit.

Or maybe you’re just so beaten into submission you’ve already given up hope for the future entirely and having this amount of money stolen from you … and likely being used against you is just another Thursday.

Too Big To Fail

OpenAI’s strategy, if the national security angle doesn’t quite pan out seems to be to try to become too big to fail. They are trying to make themselves indispensable for the economy at large. They might try to convince the government that if they fail, it will be the end for global US AI leadership. That OpenAI IS the future of the technology. They might try to convince them that because thousands of companies rely on their product that it would be an unmitigated disaster if they died.

However, they’re not really all that important. It is true that there are a bunch of companies that I would refer to as ‘wrapper companies’, whose entire existence seems to be upselling OpenAI API access. Also, many companies do use OpenAI models for various purposes. Millions of programmers use Codex and so on.

But here is what really would happen if OpenAI died:

• Enterprises would scramble to swap APIs to Google or Anthropic
• ChatGPT users would migrate to alternatives
• Some startups built on OpenAI-exclusive features would have a bad quarter
• Children would have problems cheating on their homework for a while
• Content Mills have to swap models for slop-generation

These are not problems which are catastrophic at societal scale. OpenAI is not a bank. It’s not going to ruin anyone that didn’t make a very risky investment.

The timing here is almost comical. Except for Stargate, it feels like all of these compute deals have been made within the last 2 weeks or so. They make a shitload of investments and instantly try to get bailout guarantees? They didn’t even wait a moment for people to forget that all these deals and promises have just been made. Let me ask you a rhetorical question: Would OpenAI be doing this if they were actually confident about being able to pay off these investments? If they don’t think they can, then why did they agree to it? It almost looks like being bailed out in the end was their plan from the start.

“Doomers”

I’d like to extend a special thanks to the real Gs like Eliezer Yudkowsky, Connor Leahy and Geoffrey Hinton for raising public awareness for the dangers of AI. Without their efforts, the credibility of this largest theft in human history would be in shambles, but these guys, they worked tirelessly to make sure the public will eat this up.

If OpenAI didn’t get this bailout assurance - or explicitly an assurance that they will not be bailed out, no matter what - investors would still have to fear that there is any chance of them just losing their money and they would have to think a little about investing quite as recklessly into OpenAI. But if they know in advance that bailouts are coming, well, what do they have to lose? At that point they’re just spending little 2-year-old Joannas college fund… and directly transferring it into their own bank account.

So yeah, Eliezer, Geoffrey, this would not have been possible without your efforts. Thank you so much. I understand you’re genuinely concerned about the future for this species, but your intentions have been perverted and are being used to harm everyone.

If there was a chance of just letting the bubble pop, the risk of near-term AGI would actually be way lower. I’m quite confident they can’t build AGI using the methods currently available to them. So, normally, once they failed to provide meaningful model quality gains with their increased compute budgets, they would have simply ran out of money for development and downscaled to primarily serving the models they already had at that point while keeping a smaller research division alive. Operating like a normal company, essentially.

It Gets Worse

Oh, by the way, do you truly believe that if OpenAI gets government assurances that they’ll be the only ones? What about Anthropic, Google, Meta, XAI, Amazon, Apple and Microsoft? Well, they are clearly no less important in regards to national security. So why don’t they all just start making compute deals like this? You could pretty quickly be looking at $10T or more. Are you ready to have $30k+ vanish from your bank account to fund the AGI gamble? Just remember, if they succeed, they’ll mostly just use it to eternalize the current power structures / make them worse. You’re not actually going to benefit from it. There’s like no realistic scenario where it’s actually making your life better as long as one of these maniacs has control over it.

All-in-all, this entire gamble on AGI is utterly insane. Given the >$10T total investments in the responsible companies thus far, the money is clearly there to just go fix every single issue currently negatively affecting quality of life. Why would anyone need AGI when we just need a bit of a wealth redistribution? Why gamble when you can just have the automatic victory? We could still argue about whether or not to build AGI after poverty isn’t a thing in the US.

Do Something

I am asking every US citizen to go out into the streets and demonstrate against this, or at the very least contact their representatives and demand they oppose federal guarantees for AI companies.

If it doesn’t work, riot. Topple the fucking government. I mean, all your peaceful options have already been taken from you. The media is bought by the wealthy and government. The courts are compromised. Companies write the laws via regulatory capture. Bribery is legal via lobbying. Both parties are funded by the same donors. The government is just openly doing crypto rugpulls and insider trading now. The system is itself being designed to prevent reform. The government is purposely pitting people against each other, such as with their war against transsexuals or immigrants to keep them pacified from doing anything against their real oppressors. After all, people that are divided, hopeless and alone are so much easier to control.

Seriously. Revolution is the only remaining option. Yes, you will probably either lose your life or spend the rest of it in prison labeled as a terrorist. On the other hand I will forever revere you as a hero and you could be one of the contributors to freeing hundreds of millions of people. People often look to countries like North Korea, asking themselves “Why don’t the people just revolt? How much worse would things need to be for them to do something about it?” Well, it is plainly obvious that no amount of suffering necessarily causes people to stop putting up with it. They’ll let themselves be exploited to their last breath. Though, Kim Jong Un never pulled the sort of transparently corrupt shit the US government is doing.

Some Words Of Empathy

I understand that US citizens are currently dealing with a ton of confusion, a billion issues that appear just like this one, they don’t know which issues to lend their attention to, their media is bought and fueling the culture war, the government is shut down, everyone is trying to just somehow get by.

I get it. The situation is genuinely awful for just about everyone. But there needs to be something where people stop fighting with each other and decide that complaining on social media about the situation isn’t sufficient.

This seems to be a good issue to finally show a united front against the ongoing exploitation. The political affiliation of someone really shouldn’t matter here. Nobody likes losing this much money. The total amount of money for this one incident is so high, it probably represents the largest transfer of wealth in human history. If this particular instance is not worthy of being fought back against, I’m not sure that anything would ever be. To just let this quietly happen would show that the people of the US have been defeated. That they have given up any hope in being able to resist against their exploiters and that they’ll just sit by quietly as every last penny is being extracted from them.

I do this thing, where I just have a black marker and a map and if a society makes it clear they have been defeated by their oppressive regimes, I paint it black. I try to stop thinking about the place, as imagining the horror becomes too painful. I haven’t done it for the US quite yet and I’d prefer not to. Come on USA, you can do it! You can avoid the wrath of my marker!

Conclusion

Message your representatives, go demonstrate, specifically make it be heard that you don’t want to fund OpenAI bailouts, then riot when it doesn’t work.

The “Civility” Trap

There’s this thing where if you point out genuinely insane corruption with appropriate emotion, you get dismissed as “unhinged” or “conspiratorial.” But if you’re calm and measured about it, you’re implicitly treating it as normal political disagreement rather than, you know, massive theft and erosion of democracy.

I selected the tone of this post carefully. I don’t want to normalize the greatest theft in human history.

ADDENDUM

It appears that the US public actually caught this one and was immediately outraged. This has forced a public statement that ‘There will be no bailout for AI companies’ and Sam Altman is doing damage control.

Good job.

Nonetheless, many people are not quite buying it. The whole thing could still be rebranded or one of the things I suggested the govt might do could happen (covering cost and acquiring equity or compute infrastructure themselves).

I still suggest staying vigilant of this matter.