Photos by Lars Tunbjörk , from his photobook Office
Two days ago, someone named Matt Shumer posted an essay on Twitter with the title “Something Big Is Happening.” Almost immediately, his essay went extremely viral. As of this writing, it’s been viewed about 100 million times and counting; and it’s been shared by such strikingly diverse figures as the conservative commentator Matt Walsh (“this is a really good article”) and the liberal pundit Mehdi Hasan (“perhaps the most important piece you read today, this week, this month”). I’ve heard countless reports of people being sent the article unprompted by parents and siblings and friends. I expect that Shumer’s essay will end up being the most widely-read piece of long-form writing this year.
And it’s not hard to see why it’s touched such a nerve. For most people who use it, “artificial intelligence” amounts to free-tier ChatGPT; for them AI is useful for answering questions and drafting emails. But now people are starting to realize that AI will be a massive force in the world. This is the year that ordinary people start to think about how it’ll change human life. And it shouldn’t shock you that the first thing they’re thinking about is whether AI will take away their jobs, render any skills they have worthless, and make their lives worse. There’s a growing sense of panic. The Atlantic is talking about AI job loss. Bernie Sanders is talking about AI job loss. Matt Walsh says that “AI is going to wipe out millions of jobs. It’s happening now. Everything is changing. The avalanche is already here. Most of what we’re currently arguing about will be irrelevant very soon.” We are entering a moment of panic.
So it’s the exact perfect moment for someone claiming to be part of “the AI industry” to write an essay saying that we’re in a moment just like February 2020, watching COVID infections go exponential. And just like COVID did, artificial intelligence is about to crash into ordinary people’s lives with unbelievable force; and that the only way they can get ahead of this massive shock is by buying subscriptions to AI products, saving more money, spending an hour a day experimenting with AI, and perhaps following Matt Shumer so that you can “stay current on which model is best at any given time.” It’s not a very good essay—much of it was transparently generated by AI, and Shumer admits as much—but timing and positioning are much more important factors for the success of any argument. And Shumer’s timing and positioning were impeccable. I don’t think that any other piece of writing is going to do more to shape what ordinary people think of AI than Shumer’s essay will. It’s going to become a key document of this moment.
And that’s a very bad thing. Not because the essay is AI-generated, but because it’s entirely wrong on the merits of what AI will do. I don’t think that we’re living in anything like the pre-COVID moment of February 2020. I don’t think that ordinary people have very much to worry about from AI. And I don’t think that the forecasts that people are taking away from the essay—imminent mass job loss, the entire world transforming rapidly starting in the next few months, “the avalanche is already here”—are grounded in reality. And I’m worried that these misperceptions are going to result in disaster.
I’m not saying this because I don’t believe in AI. I think that AI is going to be extremely important; I expect that it will end up being at least as important as the discovery of electricity or the steam engine, and there’s a good chance it’s the most important thing that humans ever invent. I think that the future is going to be very different from the past.
But I don’t think that this implies the “February 2020” world that people seem to imagine. I honestly don’t think that we’re going to see mass unemployment, or the sudden death of human cognitive labor, or anything that feels like an “avalanche.” The years to come will be weird, especially if you’re keeping abreast of the latest developments in AI. But the actual impacts of AI in the real world will be a lot slower and more uneven than people like Shumer seem to think. Human labor is not going away anytime soon. And whether or not they spend an hour a day using AI tools, ordinary people will be fine.
Actual labor substitution is much harder than people think
AI will be extraordinarily capable: it will continue to awe and amaze us with what it can do, and it will only get better, and it will get better at an accelerating pace. There are already a lot of tasks where AI is as capable as a competent human; and that number is only going to grow.
But that doesn’t actually imply the mass replacement of human labor.
... continue reading