Every Generation Thinks This Time Is Different: The AI Job Myth Transcript, AI Summary & Key Points

00:00 If you're young and you're trying to figure out what to do with your career right now, you're probably hearing this. AI is going to take your job. And even if you aren't so young, you're probably hearing the same thing loudly from everyone. And the honest response to that is, every generation has heard some version of it about the machine of the moment.

00:23 Cars, electricity, the PC, the internet, smartphones, every single one of these was predicted to permanently destroy work. And every single one created more total jobs than it destroyed. Just different jobs. So Jeff and I want to walk through the pattern because the pattern is real and it's consistent. But from the inside the moment, it never looks like a pattern.

00:48 It always looks like the exception. It always looks like this time is different. Which is where I want to start, because this time has been said a lot of times. All right, wave one, cars. The turn of the 20th century, one of the biggest employers in most cities was the horse economy. Oh, this is going to be my favorite segment. You're right. You're a barrel racer, so I can see why you would care about this.

01:14 I love me some fast horses which is very inconvenient for a video about the technology that replaced them. Sad. But the good news is nobody has built an AI that can turn a barrel yet. And A, I better never take the thrill of turning barrels away from me. Sorry, go on, the horse economy. Don't worry Bree, no one wants to watch a robot rodeo. So yeah, we've got stable hands, blacksmiths, carriage makers, ferriers, grooms, feed suppliers, manure haulers.

01:46 Yeah, that was a real job title, believe it or not, and sometimes sounds a lot like my job at times. New York City, around 1900, had roughly 200,000 working horses in it, producing millions of pounds of, yeah, you guessed it, manure. Every day. Which sounds like a job you'd want a machine to take. Yeah, it turns out that you definitely wanted a machine to do that.

02:12 Cars arrived, and within about 30 years, the horse economy had mostly gone. Hundreds of thousands of jobs vanished. And that's the version of the story that told the jobs that disappeared. Right. The part that doesn't get told, though, is about what happened next. We had jobs that started to appear. The auto industry created assembly line work, mechanics, gas station attendance, road construction, motels, drive-in restaurants, traffic engineers, auto insurance, driver's ed teachers, an entire economy that just didn't

02:42 exist before. More jobs than the ones that were lost. Way more. And in fact, none of them were predictable in 1900. Nobody in 1899 said, you know what the future is? Motel clerks. Because nobody has thought of a motel yet. Nobody had thought of a motel yet. That's because they lacked the vision to see what was on the other side of the mountain. They just saw...

03:18 Fast forward. The same pattern shows up over and over. Electricity kills the gas lighting industry. I think you mean street lamp lighting industry because gas lighting is a different problem. I mean, the street lamp lighting industry and the ice cutting business. And more icebreakers. That's right. But it creates the appliance industry, the entertainment industry, and eventually the entire computer industry.

03:48 All of which were much bigger than what actually got replaced. The PC in the 1980s kills the typing tool and the filing clerk, but creates software developers, systems analysts, IT support, the entire tech sector. The internet in the 90s kills travel agents, video rental stores, and encyclopedia salesmen. Well, those encyclopedias were always a little bit cumbersome to carry around after all.

04:17 But it creates e-commerce, digital marketing, UX designers, community managers, and roughly every job that ends with the word engineer and did not exist in 1995. Then smartphones after 2007 kill the standalone GPS unit and the MP3 player, but create the app economy, the gig economy, and the creator economy. It was millions of jobs. Oh yeah. And Jeff.

04:45 Here's the stat I want people to sit with. David Otter at MIT has been doing this analysis for years. His 2024 paper on new work found that more than 60% of the jobs Americans do today didn't exist in 1940. Hold on, according to my calculations, that's more than that. That's right, more than half. Those jobs weren't sitting there waiting to be discovered.

05:15 Humans invented them on top of the tools they were handed. Again, these were all waiting just on the other side of the mountain. What brings us to the next part then is what surprises people the most. Bree, when ATMs first got installed everywhere in the 70s and 80s, what were bank tellers supposed to do? Disappear, I guess? Well, that's what a lot of people were predicting at that time.

05:42 You know, you've got a machine that dispenses cash and that should be the end of tellers. Well, what actually happened was the number of U.S. Bank tellers went up from around 300,000 in the mid-70s closer to 600,000 by 2010. So economist James Besson has the numbers on this one. Because? Well, because the ATM that made a branch cheaper to run meant fewer tellers per branch, so banks opened more branches, and the job of the teller itself actually shifted.

06:18 From just counting cash to helping customers with more complicated stuff. Same title, different work, and actually more of them. This is a version of a very old economic principle, and this is your territory, Jeff. Yeah, the very old stuff is always my territory, thanks. Yeah, I'm referring specifically to a thing called Jevons Paradox, which I've touched on in a previous video of mine.

06:42 It's named after a guy, William Stanley Jevon, who wrote a book back in 1865, and no, I was not there when it was written, that was called The Coal Question. His observation was simple. When steam engines got more efficient, you didn't need as much coal to run one. So the conventional wisdom held that coal consumption would go down as a result. Plot twist, that's not what happened.

07:06 It went up. You've seen this movie, haven't you, Bree? Well, no spoilers for the folks at home. Yes, it went way up. And because the more efficient the machine, the engine, well then the cheaper it got to operate and it could do more work for you. And that meant more usage, more people found more uses for these machines and therefore coal consumption actually increased.

07:34 That's right. And the same pattern shows up with software. That's right, same pattern, AI makes writing software faster and cheaper. The naive prediction is great. Now we need fewer developers, but remember what the Jevons pattern actually predicts is that more software gets written, more things get built that were too expensive or impractical to build before and demand for developers who can now do more actually goes up.

08:02 Which is not what the headlines use. No, it turns out headlines don't do 160-year-old economics, and headline writers don't often look over the other side of the mountain to see might be waiting there. Okay, so here's the part I want people to hear. Every generation has this exact conversation. And it's with the exact same confidence. In 1930, John Maynard Keynes wrote an essay called Economic Possibilities for Our Grandchildren.

08:34 And in it, he coined the phrase technological unemployment. His argument, machines are getting so productive that within a hundred years, we'd all be working 15-hour weeks and running out of things to do. Yeah, we were supposed to all have personal jet packs by now too though, but tell me, how's that working out for you? That 15 hour work week? Not so great if you ask me.

08:58 I'm doing at least three of them every week. More is better then. So tell me about it. On second thought, don't. But you know what no company ever says? I think we're finished. We're done. Nothing else left to do. Nope, they don't say that. Nature abhors a vacuum and so do companies. They will always fill the void with, wait for it, more work. Right, but we keep forgetting that.

09:26 In the 1960s there was a presidential commission on automation because everyone was convinced the machines were about to make labor obsolete. In the 1980s it was Japanese robotics, in the 2000s, it was offshoring. Every wave had a very smart person saying this time is different. And each time the tools got adopted and humans reorganized around them.

09:53 Because the pattern is in tech eliminates jobs. The pattern is tech eliminates tasks and humans invent new jobs around what's left. I like your angle here, Bree. A job is a bundle of tasks. So you've got some tasks in the bundle that get automated. Well, the bundle regroups. Sometimes new bundles form. Sometimes people just get to do more interesting parts.

10:19 The Bayteller got to spend less time counting 20s and start actually helping people. The developer today gets to stop writing boilerplate code and start solving the actual problem. And from inside the moment, this always looks like the exception because you can see the tasks getting eliminated in real time and you cannot see the jobs that haven't been invented yet.

10:42 That's because they're not on the other side of the mountain. In 1899, nobody saw the motel clerk over here. And nobody in 2019 saw the prompt engineer. Which is a six figure job now. Which is a fifth, sixth figure job now. That's exactly the point. Somebody watching this video is going to have a job in 2040 with a title that would say, would sound completely made up if we set it out loud right now.

11:11 And they're going to be very good at it. So if you're early in your career and the headlines are telling you your future is cooked, I'm gonna let you in on a little secret. You're actually not that special because we've done this before. They're telling you what everybody's grandparents were told. The tools change what the work looks like, they don't change the fact that there's still more work to do than we have actual humans to do it.

11:36 Yep. So learn the tools. Get good at the parts of the work the tools can't do. Get comfortable inventing the jobs that don't exist yet, because that's what every generation has actually done. Even the ones that we're sure this time is gonna be different. Especially those. In fact, despite what the prophets of doom have been saying, the unemployment rate both in the US and globally has remained largely unchanged since the advent of generative AI.

12:05 So what that means is for every horror story of AI-caused layoffs, many of which were really just a way of hiding poor hiring practices, there must be a relatively equal number of Otherwise, the rate would change. So Jeff, what's on the other side of this AI mountain? Well, nobody knows for sure, but just remember this, curious people don't get replaced, they get liberated.