Why Would I Deceive You?: AI and the Tech Bubble Scam

Dispatches From A Collapsing State receives no corporate funding and no under-the-table investments. The reason this project can provide this kind of independent analysis is because of paid subscribers. If this work matters to you, if independent analysis and reportage means something to you, please consider becoming a paid subscriber

---

Over Labor Day weekend, at the Telluride Film Festival in Colorado, audiences were treated to a surprise screening of a new documentary called You Can See Everything, a nearly three-hour documentary directed by Nathan Fielder and Lance Oppenheim focusing on disgraced Theranos founder Elizabeth Holmes in the leadup to her eleven year imprisonment for fraud. Following the screening, a trailer was released and instantly became a viral sensation.

That trailer, which features an unsettling close-up of Holmes’ face as she assures Fielder she is trustworthy, quickly turned into a stand-in for someone lying while maintaining an uninterrupted veneer of sincerity. The energy is disturbing. Off-putting. Frightening, even. And, without any prompts or explanation, viewers are left to feel as if they are in the presence of someone who cannot help but lie and who does not recognize the person on the other end of the performance as human.

“I’m engaging with you as a human being,” Holmes says. “Why would I deceive you? There’s no reason for me to do that.”

“You’re being real right now?” Fielder asks.

Holmes blinks. “I’m always being real.”

Always is doing a lot of work here. Holmes’ prison sentence stems from her fraudulent tenure at Theranos, where over $700 million dollars was raised on the promise of a technology that never worked. She claimed that she would “democratize healthcare” and produce a device that, through analyzing a single drop of blood, could reveal the secrets of the body and diagnose any number of diseases. It was never possible, but that didn’t stop Holmes or Theranos from faking their results and surrounding themselves with powerful figures like Henry Kissinger and General James Mattis. Until the walls closed in, it never mattered that the promise was a lie or that the results were faked. All that mattered was that the lie was believed.

Holmes and Theranos are held up as aberrations of the tech industry, but the horrific truth is that they were the norm. Tech is a confidence game. We are supposed to believe that these firms and projects are helmed by the smartest and most talented individuals the world has to offer. They produce glittering products while promising they are revolutions that can fit into our pockets or simply float invisibly through the air. Every product reveal splits history in two: the moment before and the moment after. In exchange, they are gifted billions and trillions of dollars. There is no oversight, merely happy, clueless complicity and economies and politics that are tuned to their every whim.

To keep the bountiful grift going, we are promised everything. Miraculous medical cures. Rocket trips to the stars and immortality. Thinking machines that will solve all of our problems if only they are kept from destroying us.

And, so, forgive me for not buying it when a completely unknown individual claiming to have resigned from Anthropic pops onto the timeline and sets social media aflame with these posts.

Coxon’s “warnings” immediately went viral and have now earned nearly 200 million views. Anthropic’s CEO Dario Amodei was quick to belch forth his agreement and was soon followed by OpenAI’s Sam Altman, Elon Musk, and much of the tech sphere. Within a day, former president Barack Obama was joining the chorus and corporate media was awash with one segment after another about the dangers of an “AI apocalypse” and the need for development to “slow down.”

The crux was that a sentient, self-improving AI could infiltrate all parts of our lives and, through one reason or another, extinguish human life. Very scary. Save for the fact that AI is not sentient, it isn’t self-improving, and any existential threat has yet to emerge.

Quickly, I need to say, I do think AI is both reprehensible and dangerous. There are sectors of work, including in the medical field, where there are benefits, but this project is wrongheaded, directed by unethical criminals, and has already had disastrous consequences. There’s no counting how many people have already died, whether through AI psychosis or from military operations, which includes war crimes around the world. Crimes have already been committed, ranging from the production of sexual abuse materials to the recent Hugging Face hack, but these crimes have been committed because the people directing the projects have either allowed them to be committed or weren’t ethical and capable enough to stop them. Hugging Face, in particular, has caused huge waves because it sounds incredible and dangerous until you realize the agents that carried out the hack were directed to do just what they did. This wasn’t a rogue AI. It was AI doing what it was told to do.

In a world that made sense, this industry would be regulated out of existence and the people in charge of it would be prosecuted and jailed. We do not, however, live in that world.

It might seem counterproductive, but the AI slowdown panic is to the benefit of the AI companies themselves. What is being sold to us is yet another Theranos, something giant and revolutionary. Dangerous, sure, but what world-changing inventions aren’t dangerous? We’re supposed to believe that these firms have cracked the code, so to speak, and have a technology on their hands that is nearly unfathomable in terms of its potential. Never mind, of course, that what they have are large language models (LLMs) that merely guess, and often guess wrongly, as to what the user will want to read or see.

In essence, these firms and their dipshit mascots are performing fear in hopes you’ll believe that they have essentially split the atom. The potential, and this is any con-artist’s secret weapon, is only as large as your imagination.

But the truth is much more mundane, even as it is still disturbing. The AI firms have reached the limits of their con. They have inflated an economic bubble to the point of popping. Their entanglements with investors have left them at the edge of a cliff. In order to reach this point, they have been pumped by hundreds of billions of dollars of investment by the other major tech firms and through loans and investments based on estimates of future earnings. There’s a problem, though. Neither Anthropic or OpenAI have figured out how to make money. In fact, they’ve already been lapped by free versions of AI programs and the corporations that bought into AI have already been shocked by astronomical bills (one business was allegedly presented with a $500 million dollar bill from Anthropic) and a piss-poor return on investment. Some have already trashed the AI that was supposed to “revolutionize” their businesses and rehired those they laid off.

Meanwhile, the AI firms have a reckoning coming their way. Unless they make good on their promises and supply massive computing responsibilities they simply cannot meet, the bubble will likely pop and their firms will die. And don’t forget that that bubble represents a significant portion of the American economy and is a series of dominoes that are already teetering on the edge of collapsing.

So the confidence game must continue. Faced with economic annihilation and the failure of the AI grift, it seems the industry is moving onto the next lie. There needs to be a reason for the development to slow down while not spooking investors and the failure to meet expectations needs transformed into a noble decision rather than an innovative boondoggle. They have not created autonomous recursive self-improvement, but what if they did? Wouldn’t that be frightening? It is yet another fantasy, this one a nightmare, that is meant to shift the focus from the previous fantasy - that they would meet their expectations and then some - while mitigating the fallout.

In other words, Theranos reaching the point of development where people need results from tests and so they simply fake them.

There are other avenues here because this level of fraud isn’t simple. AI is being cast as an essential and existential battle for future dominance. Like the aforementioned development of nuclear weapons, the AI firms and the tech world have successfully sold the concept that the Cold War between the U.S. and China will be won by the country that masters AI. And, that might be true when history plays out. We will see. But the new fearmongering around the calamitous potential of AI casts an eye to a future that doesn’t exist yet and opens the door to the state rescuing this failure of a project.

If AI is the new nuclear weapon, then the U.S. has a responsibility to make it succeed.

If the free market can’t support that endeavor, then public rescue is necessary.

If China is directing their development as a state, why shouldn’t we?

So, if all this is “true,” then shouldn’t our public resources be directed to making sure the government and AI are integrated as closely as possible and these firms are encouraged to win the AI arms race under the watchful eye of government regulation? Never mind, of course, that government regulation doesn’t exist, the firms are directed by liars who have no actual concern about the fate of humanity, and that the nightmare scenarios being peddled have no basis in reality.

The coverage and discussions around this matter are split. In the popular imagination, the public is fevered with visions of sci-fi movies like The Terminator or The Matrix. Coverage reads like a poorly updated new edition of the Cold War’s fears over space dominance and nuclear proliferation. It is a rather predictable and pitiful condition.

And a look at the political landscape reveals the rest of the picture. Both parties are awash in AI and tech money. There is no party-level opposition to these projects or the construction of data centers that pollute and steal and lessen the quality of life for everyone in their vicinity. You have Trump telling people they’re stupid if they oppose them and the Democratic Party shrugging and telling you they’re inevitable. There is no body presently in power, no representatives at the switches, that understand any of this let alone who are capable of overseeing regulation in a way that would matter or be more than yet another historic transfer of wealth.

We must view this massively important issue with clarity and refuse to be swayed by fearmongering fantasies that only make us act out of activation and fear. AI should be regulated and largely opposed. But the answer is not to bring these criminals into the bosom of the state, which has already largely been taken over by them and their billionaire friends already, and give them the keys to the kingdom. They have no split the atom and, even if they had, this is no reason to listen to them or follow them or entrust the future to them. They have already revealed themselves not to be trustworthy or even worthy in general.

Instead, we should view them as they are. Unblinking liars who have built their fortunes and power bases on lying as they have hurt untold numbers of people. When they speak, it is right to hang on every word in search of what the truth is underneath the mountain of lies. It is right not to trust them. And, again, if this world made sense, we would already be preparing them to spend at least a significant portion of their lives behind bars.