Jacob Coxon: how to run an AI doomsday media campaign
The AI doomsday is the idea that one day, an artificial intelligence will get so smart it takes over the world! And it’ll kill all the humans, just as a side effect! Pretty scary, huh. You might be familiar with this idea because OpenAI and Anthropic use it in their marketing — all the time. The…
The AI doomsday is the idea that one day, an artificial intelligence will get so smart it takes over the world! And it’ll kill all the humans, just as a side effect! Pretty scary, huh. You might be familiar with this idea because OpenAI and Anthropic use it in their marketing — all the time. The chatbot could kill us all! So it’s definitely powerful enough to replace all your employees. We’ve spoken previously of the subculture known as effective altruism. Which is a nice name, but what they do is they push AI doomsday. EA has a ton of money from weird tech billionaires. So EA sets up fake anti-AI astroturf groups talking about AI doom instead of real AI problems. They pay YouTube influencers to suddenly talk about AI doom — as if this is real science and not weird cult stuff. They do waterboard the influencers, though, so that’s something. On Wednesday evening, Anthropic AI researcher Jacob Coxon posted a thread on Twitter that took off like a rocket: [Twitter, archive] I resigned from Anthropic today. I spent the last three years doing pretraining research at both OpenAI and Anthropic. Neither company is acting responsibly. They are racing straight to self-improving superintelligence and gambling with our lives. The thread hit the media immediately. Specifically, it hit the Wall Street Journal 18 minutes before the first tweet went up. [WSJ, archive 10 September, archive 8 September] Coxon’s thread has the only tweets on this fresh, paid Twitter account. A bunch of Anthropic employees retweeted Coxon within minutes, saying how true and important the AI doomsday was. This whole thing was an organised promotional campaign. It looks like obvious fresh-cut astroturf because it is. By the way, the chatbot cannot possibly do any of this. If the chatbot destroys humanity, it’ll be from climate change. The AI doomsday is a distraction from the real problems with the chatbot. The AI vendors want you talking about Skynet — and not talking about the data centre down the road. Every report of the chatbot breaking out of human control turns out to be because the researchers told it to. Anthropic especially. Every time. Coxon is hard to find online — he doesn’t do social media, and it looks like he purged his old socials ahead of this campaign. But he’s a real guy. He did mathematics at Trinity College at Cambridge. Before Anthropic, he was a researcher at OpenAI. He worked on GPT-4o. Coxon has long been fully into the rationalist and effective altruist subcultures. He believes in the AI 2027 doomsday report. He told Wired this was “crunch time for humanity.” That’s a phrase rationalist founder Eliezer Yudkowsky has been pushing for 20 years. [Wired] The Coxon campaign appears to be a service of ControlAI, funded by Jaan Tallinn, who made his money with Skype. Tallinn discovered Yudkowsky in 2009 and he went full AI doomsday. He’s also an Anthropic shareholder. The media campaign is going great guns in just two days. Coxon is doing TV and newspapers at a fantastic pace. A pile of US and UK politicians are citing Coxon’s tweets to say it’s time to regulate AI. But only in ways that solve imaginary problems, not the real problems. The trouble with AI doomsday marketing is that if you take these guys seriously, then we need to ban the chatbots immediately. We need to shut down Anthropic and OpenAI. We need to put Altman and Amodei in jail. We’re not doing that because none of this is serious. This is complete nonsense from cranks. And everyone sort of knows that. Well, maybe we could put Altman and Amodei in jail for everything else. The goal of this campaign is to get a regulatory moat around the chatbot industry — regulations that only the big guys can meet. And not do anything about the actual problems with AI. This is some of the widest publicity AI doom’s had for a while. Here’s to it giving the Anthropic IPO that nice final push. Video — PodcastSource: Pivot to AI — Published — Category: Business