I will get into the issue of AI doomerism, AGI hype, and why we don’t need to panic about the AI apocalypse in a second. But one brief side tangent first. 

I am currently writing this newsletter from the airport in São Paulo, Brazil, where I’m on my way to a friend’s wedding! I’ve never been to Brazil before, but I have now been lucky enough to visit three countries in South America (the other two being Colombia and Argentina). There’s a booming tourism industry in places like Cartagena, Buenos Aires, and Rio de Janeiro, and just this week an excellent piece about Western misperceptions of Brazilian culture by Nicole Frio for The Flytrap landed in my inbox. It’s Brazil-specific, making it even more fitting for me this weekend, but I think a lot of what Nicole is talking about here could apply more broadly to cultural distortion. Something to think about!

Now back to the topic at hand, which starts with a viral X thread posted on Tuesday by Jacob Coxon, a former AI researcher at Anthropic who reportedly worked at the company for a grand total of about six weeks (or at most a few months). He posted that he was resigning from the role over fears that big AI companies are racing to create technology that has a good chance of killing us all in the next few years (via, of course, unspecified means). 

Now, this kind of bold apocalyptic proclamation is nothing new from the AI industry. It’s a tried-and-true way to build market hype and influence the trajectory of government regulations based on literally nothing. But it also has the effect of terrorizing the public. When I reposted a skeptical Bluesky thread about the whole thing on my Instagram Stories, I got a flood of messages from anxious folks and people who don’t understand the perceived benefit of saying your industry is working toward human extinction. Which is fair! So I wanted to break down further where this messaging is coming from, why the biggest AI companies love to push it, and how they can weaponize it to their advantage while the rest of us suffer from AI pollution, AI nudification, AI disinformation, and more corrosive aspects of their so-called advances.