A post about AI made to X this week sparked political and media firestorm. In fact, it currently has over 158 million views:
The thing that seems to have really caught people’s attention was when Coxon said, “The people building AI earnestly believe that it could kill us all by the end of the decade.”
Of course, as anyone who regularly reads Culturcidal knows, this sentiment is not new. In fact, more than a year ago, in an article called “Can We Stop AI from Destroying the Future?” I wrote:
“However, one of the biggest goals of people working with AI – and it’s impossible to say at this point if this is something that will happen in a few years or never – is Advanced General Intelligence (AGI). When people talk about the potential of AI ending the human race, this is the danger they’re talking about… Some people don’t think this is a serious concern. Those people are foolish and naïve.”
In that same article, I linked to Elon Musk’s famous interview with Joe Rogan in 2024. Apparently, many people have forgotten this, but during that interview, Musk openly said there was a 20% chance that AI would annihilate us:
Why is this sentiment, which has been floating around in circles that are familiar with AI for years, suddenly catching fire now? Good question, but people unfamiliar with AI are grappling with a lot of good questions right now, like…
* Why are the AI companies moving at such breakneck speed?
* If AI is extremely dangerous (and it is), why are we making it at all?
* Why do we need so many data centers?
In this article, we’ll answer those questions, but to do that, we’ll need to go all the way back to WWII and the development of the nuclear bomb.
Germany started working on nukes first and America had to play catch-up. Fortunately, for a variety of reasons, we did win the race to build nuclear weapons. That meant we had and used a nuclear bomb first. If Germany had beaten us to the punch, we might all be living in an eternal Nazi state.
Incidentally, setting off a nuke was not risk-free. There were serious discussions about whether a nuclear bomb might ignite the atmosphere and wipe out all life on Earth, although the scientists had concluded that wouldn’t happen when they tested it… probably.
So, what does that have to do with AI?
Well, AI is the most revolutionary technology ever created, and once again, two countries are ahead of the pack in developing it: the United States and China. Just as the world would look very different today if Hitler had beaten us to a nuke, the world will look very different if China beats us on AI.
Doing so would allow them to pass us economically and militarily and replace us as the world’s dominant superpower. The people who run China are not good people. They believe in harsh, stifling, top-down control, and the world under China’s rule could be just as grim over the long term as a world ruled by the Third Reich.
If the United States wants to keep its place in the world and the world wants to remain free, beating China on AI is not optional; it’s a necessity.
This is where the data centers come in. There are already several thousand of them across the United States, and these AI companies are DESPERATELY trying to build more for the same basic reasons phone companies relentlessly laid fiber optic cable everywhere in the early days of the Internet. Getting “more compute” lets AI advance faster, and data centers are what let AI get “more compute.”
The problem with that has been twofold: First, to gain an edge in the AI race, China has run influence operations in the US to make data centers unpopular, and they’ve had some success. Additionally, there have also been some REAL problems with data centers in the past. Most use a lot of power, some use a lot of water, and others are loud.
Are these fixable problems? Yes – and increasingly, as American public opinion has gotten more hostile to data centers, they are getting fixed in a way that makes data centers a huge plus for areas they move into:
Still, some people may be asking: If it’s dangerous, why can’t we just stop working on AI?
The answer is that the genie is already out of the bottle. If we stop working on AI, that doesn’t mean China will stop. Right now, they are behind us, but not WAY behind us. Would they suddenly launch so far ahead of us that we couldn’t catch up if we stopped working on it for a year or two? HOPEFULLY NOT, but we don’t know for sure. Especially since if the best talent can’t work on AI here and China is willing to pay a premium, how many of them might go to work for Chinese firms? We don’t know.
The real solution to this takes us right back to nuclear weapons. Over time, although there are an enormous number of countries with the resources and capability to develop nuclear weapons, only a small number of them have done so in the last 80 years.
We could do something similar with AI, and when our country and China are far ahead of everyone else, there’s every reason to think we could reach an agreement because, although both of us want to win the AI race, we also share many interests. Neither country wants nut-jobs to develop bioweapons with AI, rogue AGI getting out of control, cyberattacks exposing our government secrets, or to risk the future of the human race. Could we agree to some verifiable safety protocols that would dramatically lessen the chances of AI wiping out our species? You’d have to think so… and this is the way.
AI is here, and NO MATTER WHAT WE DO, it’s going to keep advancing and changing things. We can’t predict how all of that will play out at this early stage, but we can at least make moves that lessen the danger to the whole human race.
Good article John. Well done covering the big picture view with objectivity and balance.
The genesis of AI is really an interesting story. An open-source project from Google software engineers working on search term predictability. The code they developed was not really very big... but it surprised them in how it seemed to be thinking on its own... coming up with results that they never anticipated and did not program in.
By the time they knew they had stumbled on a working AI model, the open source was already in the hands of other software engineers all over the world.
All the assholes that pushed globalism for corporate profit maximization, they can go eff themselves for how they have put the US in a dangerous geopolitical situation.
I don't think we will all have to die for AI, I think we are all at risk of death from globalism. Because otherwise, like nuclear weapons, the US would have led and controlled AI development and thus established the rules that keep power in check and prevent the extreme race to the bottom that would result in chaos and destruction.
US companies trained China as the threat that we might die for. AI is only a piece of that giant puzzle of crap.
I've worked in IT all my life. AI is certainly a useful tool but it still has a lot of limitations. Right now it just seems like a souped up adding machine. But there is a large resistance movement growing against data centers. The biggest problem is with water usage and water pollution.