Why E/Acc Feels Psychopathic

And why doomerism feels righteous

“Applications are now open for PLZDONTKILLUS!”

This was posted by Aella on X this week. If you are worried about the dangers of AI and you like to make videos, you can join an all-expenses-paid residency for one month. Check it out at plzdontkillus.com.

It sounds like a fun opportunity to me, but not everyone is impressed. Marc Andreessen commented:

The inner party of AI doomerism is a Berkeley lifestyle grift fueled by unaccountable dark money. The outer party of AI doomerism is young people who believe the propaganda and ruin their lives.

Physicist Guillaume Verdon (better known as Beff Jezos) commented:

The Doomers are creating UGC psyop factories. The memetic warfare is reaching all time highs. This is evil and will cause mass anxiety that is more dangerous than the AI products themselves

These posts illustrate the growing animosity between the AI accelerationists and the doomers. Despite the strong rebukes of Andreessen and Verdon, a casual observer, no doubt, would side with Aella and the PLZDONTKILLUS crowd. The polling on this is clear. The public is significantly more concerned about AI than they are excited by it.

It could just be that the public is misinformed. This is understandable, given the reality that news stories about scary things spread faster than stories about happy things. If it bleeds it leads. So naturally most content that casual observers are going to see about AI will be stories tinged with doomerism.

But I think there’s another explanation that goes deeper. To put it bluntly, AI accelerationism feels vaguely psychopathic. Even if people can’t fully articulate the feeling, it’s there in the background in the e/acc vs. doomer debate.

Let me explain.

AI promises to be the most powerful invention in history. Every time a technology like this comes along, it causes extraordinary disruption and extraordinary harm. Gunpowder, rifles, the railroad, the telegraph—each of these inventions changed the world and had a direct impact on the nature of war. The printing press caused multiple religious wars. Deep-water ships contributed to the rise of empires, the destruction of indigenous cultures, and the slave trade. And nuclear weapons, of course, wiped out two entire cities.

Given this reality, there is no world where AI continues to accelerate and doesn’t lead to incredible amounts of death and misery. What exactly could this mean? Will we lose another two cities? Will we lose ten? Will all future wars become far more deadly? Will there be an AI-created virus that’s five times as deadly as COVID? Will mass surveillance strip us of all our privacy? Will swaths of the population become unemployed? Will cyberattacks take down airplanes and cause cities to go dark?

And these scenarios assume we’re on the good timeline. Remember, we’re on the good timeline for nuclear weapons. Nukes came about as an existential threat. After they were dropped on two cities, they have since been used strictly as a deterrent, ushering in an unprecedented era of peace and stability.

To put a fine point on this: Even if AI brings us every conceivable benefit that has been promised (the end of disease, a shorter work week, global prosperity, solutions for climate change, etc.) still we can expect significant harms, including hundreds of thousands of deaths, as a direct result of AI in the interim.

Accelerationists may not admit this, but they must know it. If AI is as powerful as people claim, it will cause death and misery even in the best-case scenarios. Promoting AI advancement with little consideration for the consequences gives implicit approval of those consequence.

And that’s why accelerationists come off as a little psychopathic. And it’s why Aella (as well as a lot of normies) feel like they need to plead with the AI companies to PLEASE DON’T KILL US.

This isn’t to say that accelerationists are wrong. If I had to pick a side, I’d join the accelerationists over the doomers every time.

But accelerationists do need to grapple with the fact that their cavalier enthusiasm for progress no matter the cost strikes people as unhinged.

Meanwhile, there’s a sense of safety in the status quo that AI doomers cling to. This is why they come off as comparatively righteous, even if their objective to throttle AI progress may lead to more mass graves in the long run.