The Brooklyn Rail

The leaders of the AI boom want you to think this is a wonderful thing for the world, but it’s already looking bad and we haven’t even gotten to the stage where the machines are more intelligent than humans. These AI entrepreneurs all have a quasi-religious belief in AGI and its power to transform the world and solve all our problems. They believe in AI as an Everything Machine.

Whatever the Everything Machine eventually turns out to be, it is already having an overwhelming impact on the economy. This year alone, spending on AI represents 43% of the entire growth in GDP in the US, and AI is responsible for 80% of the gains in the stock market. Nvidia alone (now worth $5 trillion) comprises 8% of the S&P 500, and Nvidia, Apple, Microsoft, Meta, Amazon, Alphabet, and Tesla combined make up more than one third of the entire index.

Each of the AI leaders are creating their own AI worlds, building huge data centers as fast as they can around the world. These data centers suck unbelievable amounts of electrical energy to run the computers and massive amounts of water to cool the computers down. And most of the energy to fuel the data centers is coming from burning natural gas and coal. Nuclear plants can’t be built fast enough for the increased demand, and solar and wind power can’t do it because of intermittency and the high cost and space requirements of energy collection and storage.

This small group of technology companies is collecting billions in profits and spending wildly to develop an unproven technology that will have to deliver tremendous returns pretty quickly for all of this to pay off. What if it doesn’t? In the short term, if the AI bubble bursts, there could be a catastrophic market correction or collapse. In the longer term, the effects of AI on human work are uncertain.

The Trump regime is all in on AI. On his first day in office, President Trump signed an executive order eliminating safety testing rules for AI, and in July, he announced that he would remove all safeguards and oversight from AI development, opening up the possibility of the worst outcomes in societal transformation. Trump also removed what few environmental regulations there were on the construction of new data centers and, of course, mandated that AI development must also remove all mentions of “diversity, equity, and inclusion,” “climate change,” and “misinformation” in their plans. “The American people do not want woke Marxist lunacy in the AI models,” trumpeted Trump.

These actions halted and reversed all previous efforts by tech entrepreneurs to convince the government to regulate their industry. In September 2023, a dozen of these leaders met at a summit on Capitol Hill and basically begged legislators to provide some oversight. At that time, Elon Musk was warning of the “civilizational risks” of AI.

But that’s all in the past. Now, the leaders of the AI industry take it for granted that they are going to take over, and the downsides and existential dangers need to be reconfigured as costs of production. The first two downsides—increasing energy prices and job loss—are already in play.

After Stephen Witt’s article on the loud, hot, dirty, energy-sucking insides of data centers was published in the New Yorker on October 27, Witt was asked,

What do you say to people who feel stressed out about all this?

I am also stressed about this. I mean, I go back and forth. The end goal here is that most of what humans do becomes obsolete.

Do you really think that?

Yeah, one hundred per cent. I think that in the future, all forms of labor will at least be conceptually done by a computer. With the combined push for robotics and hyper-intelligent computing systems, what’s left? I guess we should all go to clown school—study live theatre, or something.

Writing, maybe?

Writing, I don’t think so. No, I think the computer will catch up to us if it has not done so already.

You sound pretty confident about AI getting better and replacing us all.

That’s data-driven. The premise of all of this is that putting more Nvidia microchips in the barn will result in better AI. Empirically, so far, that has been true. Now, as the AI pioneer Demis Hassabis has wondered, how long will this work? Will we hit a brick wall? No one knows. But right now, the evidence shows that this is working.

It's definitely working for the time being, but the likelihood of hitting a brick wall is increasing, not just in terms of energy but also in terms of actual data. As Witt himself says in the article,

There is now talk of a data shortage. There are thought to be about four hundred trillion words on the indexed internet, but, as the OpenAI co-founder Andrej Karpathy has noted, much of that is “total garbage.” High-quality text is harder to find. If trends continue, researchers say, AI developers could exhaust the usable supply of human text between 2026 and 2032. Since AI chatbots are recycling existing work, they rely on cliché, and their phrasing grows stale quickly. It’s difficult to get fresh, high-quality writing out of them—I have tried.¹

What is becoming more and more clear is that the rush into AI has already completely outstripped reasonable efforts to think about what effects it will have on privacy, personal liberty, the difference between real and fake, and life and death in warfare.

In an interview Ross Douthat of the Times did with the Chief Technology Officer of Palantir, Shyam Sankar, on October 30, about mass surveillance and the future of war under AI, came this exchange:

Ross Douthat: So if we walk outside of this room and enter Midtown Manhattan, we are under constant surveillance. It’s not all government surveillance, but there’s a relaxation that you can feel where you’re like: OK, but all of this surveillance is distributed across so many different public and private entities, and unless I am literally a terrorist, the odds that people are going to be constantly watching and scrutinizing me are very low.

But then the fear becomes: Well, if we have this incredible way to make it all more and more and more efficient, then maybe privacy does start to disappear.

I don’t know. What do you make of that?

Shyam Sankar: There’s two thoughts there. One is: Well, are you saying that you feel safer because the institutions that are supposed to protect you are structurally incompetent?

And that’s the part where I feel like—

Douthat: The answer might be: Yes, sometimes—right?

Sankar: Yeah. And then a consequence of that—which I think a democracy can decide—is that they also can’t do their job. They can’t protect you from the things that they’re supposed to protect you from.

So I’d offer another solution to this, which is: They should be really good at doing what they’re doing, and we should have a strong ability to oversee that they’re not doing things that they’re not supposed to be doing.

That’s exactly what we designed Palantir to do.

And later, Sankar says,

I think a lot of the policies, a lot of things that people are struggling with right now in the US were voted on at the ballot box. What ICE is doing was voted on at the ballot box.

But “what ICE is doing” is no longer limited to apprehending criminals entering the US illegally, and the line between ICE and the National Guard is no longer clear.

On October 8, Major General Ronald Burkett, director of operations for the National Guard bureau at the Pentagon, ordered the National Guard in all fifty states to train 500 guard personnel as quick reaction forces for “quelling civil disturbances.” This would result in 23,500 troops in every state and territory on the ready at the beginning of 2026.

At the same time, Trump and Stephen Miller are reassigning ICE officers and replacing them with more aggressive officers from CBP (Customs and Border Patrol) in Los Angeles, San Diego, Phoenix, Denver, Portland, Philadelphia, El Paso, and New Orleans to increase arrests from the current rate of 900 a day to a minimum of 3000 arrests a day. They are expecting that tripling arrests will increase resistance from citizens.

As Aaron Glantz reported for The Guardian:

Janessa Goldbeck, a former US Marine Corps captain and chief executive of the Vet Voice Foundation, a non-profit advocacy group, said the order represented “an attempt by the president to normalize a national, militarized police force.”

She predicted that force would be used to send troops to states led by Democratic governors without their permission and could be used to suppress turnout and disrupt the fair operation of elections.

In a worst-case scenario, she said, “the president could declare a state of emergency and say that elections are rigged and use allegations of voter fraud to seize the ballots of secure voting centers.”²