You probably don’t know Matt Shumer’s name, but there’s a pretty good chance you’re familiar with his thoughts about AI. On Tuesday, Shumer published an essay to X titled “Something Big Is Coming,” which almost immediately caught fire online. (According to X’s not-always-reliable metrics, it stands at 73 million views as of Thursday morning.) In it, Shumer, the founder of an AI company, warns that enormous advances in technology are poised to reshape society much more quickly than most people realize. He analogizes artificial intelligence’s rapid improvement in recent months to the beginning of the COVID pandemic — a looming, seismic societal change that only a small faction is really paying attention to. And he warns that the tech sector is the canary in the AI coal mine: “We’re not making predictions. We’re telling you what already occurred in our own jobs, and warning you that you’re next.”

As Shumer’s post ricocheted around the internet, it drew a predictably divided response. Some saw it as an incisive warning of things to come, while others dismissed it as another piece of disposable AI hype or a naked money grab. I caught up with Shumer on Thursday to discuss the overwhelming response to his essay, why he used AI to help write it, and whether all of our jobs are actually in immediate danger.

Your essay has been making the rounds in a way that few things do — it broke social-media containment. What has the reaction been like?
It’s insane, because I didn’t expect this. I didn’t expect anything close to this. I originally wrote it for my mom and dad because — I was home with them this weekend for the Super Bowl — I’m 26 and I was trying to explain to them what was going on. I felt that there was an inflection point when GPT-5.3-Codex came out. I tried it and was like, Oh my God, this is not just a step better, this is massively better, and it’s the first sign of something a little scary.

The way I view it, AI labs have focused on training models that are really good at writing code, and that’s really important because what we see in the engineering space is like a year ahead of what everybody else has access to. So a model today is, let’s say, at level 50 at writing code, but level 20 at law or something of the sort. The next model will be probably at level 100 on code and level 50 on law. It woke me up, and I felt that I had to share it. I was looking around and I was like, What can I give to my parents to help them understand this, so that they’re not just thinking their idiot son — that’s probably a terrible way of putting it, but you know what I mean — is saying “This is happening” and they have no way to know what or not to believe it? 

There are a lot of pieces of great writing in this space, but they’re all extremely technical, and I think that’s part of the reason people don’t understand what’s coming. They’re written for nerds by nerds. They almost take pride in sounding as smart as possible. So I figured it would probably be important to write something that they could understand. And as I wrote it, I realized it could actually help other people. I decided to post it and it quickly broke containment. I have friends who are very much outside of the tech bubble, and it’s being passed around their offices, and they’re texting me and it’s a surreal experience. But I’m glad it’s happening. I didn’t expect my article to be the thing that did it, but it needed to happen. People need to understand what’s coming. It may not affect them today, it may not affect them in a year, but at some point it will, and I’d rather people be aware and have the opportunity to prepare than just be blindsided by forces that they can’t really control.

Did you use AI to write any of it?
I did. I actually posted a little bit about this because I think it’s important for people to know. People are responding like, “Look, this is AI-written. You should ignore it.” And it’s not entirely AI-written. I used it to help edit to sort of iterate my ideas and the ways of phrasing things, and it was incredibly helpful, but that’s kind of the point. If this was helped by AI and got millions of views, it’s clearly good enough.

I didn’t say, “Go write this article.” What I did was feed it a bunch of articles that I have read over the years that I think articulate these points really well. I said, “Here’s what I agree with, here’s what I disagree with. Here’s my unique spin and take on this.” And then I said, “Interview me — ask me dozens of questions.” And I spent over an hour answering those first wave of questions. Then we repeated it, and basically I ended up building this huge dossier of everything I believe, and everything I wanted to explain. And then for each thing, asked, “How can I........

© Daily Intelligencer