You Ask, I Answer: Jobs Lost to AI?

In today's episode, Phil asks, I am unclear why jobs will not be lost in the content revolution you describe. It is not just someone skilled in AI will replace someone who's not, it's someone skilled in AI who can quickly do work that might have involved multiple people previously, surely. Okay, so this is true, sort of. And here's what I mean. Anytime you have a major technological change, you have labor market distortions, right?

Um, you have all sorts of uh ripple effects. However, when you remove a bottleneck from one part of a process, that doesn't mean the whole process gets better unless that process is literally a single step. Most of the time what happens is that you then get a bottleneck somewhere else in the process. So, for example, let's go way, way, way, way back in time to 1764 and the spinning jenny, which is a uh a textile spinning system that allowed a worker to spin multiple threads into into fibers much faster. They could produce yarn really, really fast.

So this traditional spinners, you know, the old Cinderella style, you know, the spinning wheel and and stuff like that. The spinning jenny did a did that much, much faster. So this removed the job of traditional spinners. Um you you needed far fewer of them. But what did this do to the rest of that supply chain?

It dramatically increased the demand from the number of weavers who had to take all this stuff and turn it into cloth, right? Because you go from from fibers to to thread and yarn to cloth. And so you now have this this massive bulk of extra yarn being created by these these machines, and now you have to you have a bottleneck at the at the weaving side of things. And then, you know, 20 years later, uh, the power loom was invented, and that automated weaving. Of course, well, what does that mean?

Now you have the ability to create more cloth, and that means you can produce textiles much faster, which now means you need more people to s to distribute and sell stuff because it creates these distortions in the labor market. AI will have similar effects, right? Any technological change has huge societal effects. Right? The smartphone dramatically changed the telecommunications industry, right?

How many people used to be had jobs for maintaining public telephone booths, right? Those jobs are gone. There's like booths now that uh you know they don't they don't need to I don't can't. I'm trying to tell you the last time I actually saw a operational telephone booth, and I want to say I might have seen one in London as a more of a historical curiosity, and or it was something with Doctor Who, one of the two. But there are just they those things just don't exist anymore.

Are there still jobs for horse and buggy drivers? Yes, but not many, right? They're largely tourist attractions. Go to a major city, there's a horse and buggy driver who will take you on a carriage ride, uh, you know, horseback carriage ride around the city for tourism purposes. It is not a primary form of transportation.

When those jobs get lost, other jobs tend to appear elsewhere in the pipeline in the supply chain until you get to a point where machines are doing the entire supply chain, including the demand side, you will still have bottlenecks. And where those bottlenecks occur occur, you will have increased demand for those workers. Let's say you roll out Chat GPT in your organization, and you say, we're gonna now go from uh 10 blog posts a month to 10 blog posts a day. Great. But you also say, but we're gonna make sure that it's correct, right?

We're not just gonna let the machines foam at the mouth. So we're gonna need people to edit these things. Well, your one editor who was fine working on 10 blog posts a month, and now is like, uh, I can't edit 10 blog posts a day, guys. Um, so all those folks who are on the content team who were writers, we need to either upskill them into editors, which presumably that would be a pretty easy transition, or we need to hire more editors and maybe let the content, the writers go so that we can have more editors so that to address this block now in the supply chain. I don't see, I don't see massive amounts of um just lost jobs with nothing to replace them.

I do see plenty of jobs where yes, that job will go away, or the a large portion of that job will go away, but you will then have the supply chain constraints through at the rest of the the um uh pipeline. Think about what's happening now with um with fine-tuning of large language models, the ability to get a model to be tuned to do what exactly what you want it to do, and the innovations that are happening as of the day of this recording with like local document stores that you can use to tune a model. The job of a model content curator does not exist yet. But if this particular style of implementation takes off, and I think there's a good chance it will, because it's it's faster and easier than full you know, supervised fine-tunes of models, then that people are going to need to do that job, and there will be a strong demand for that job for a couple of years until something comes along to automate that and and so on and so forth. Again, any place you've got humans, you're gonna have trouble scaling, right?

People in general like to do business with other people, not all the time, and certainly uh for simpler transactions, people would prefer not to deal with other people, right? You just want to go in, um, you know, press a couple buttons, get your driver's license and leave and not have to wait 45 minutes and you know drink steel coffee. Uh that's not a fun experience. But until machines are making decisions and doing purchases and stuff in the supply chain for information and knowledge, we're still gonna need people. And in fact, we're gonna need people probably more so than pro previously, because we've got to deal with the increased demand.

Think about farming, for example. Right? Farming used to employ thousands of people per farm to pick produce, to uh inspect it, to to package it, to get it to market. Now, a farm has far fewer of those people, right? Now farms have workers that uh, depending on the crop, uh are driving huge machines around.

And these huge machines are processing the the goods and and getting them ready. Well, now you need people to handle the increased output of the farm, and the market itself is continuing to grow because the population of the world keeps getting bigger. And so there's even more demand for jobs downstream. You do you need a thousand people picking corn anymore? No.

You can have one industrial uh combine that can do that really well. But do you still need people to get it inspected, cleaned, shipped to the store, etc.? Yep. You need more of them than ever. So the watchword for people in their careers is agility.

Uh do you have the agility and flexibility to change how you do business and your role within a business? If you do, you're gonna be fine, right? In fact, you're probably gonna be more valuable than ever. If you don't, you're less likely to be fine. And it's not gonna be one of those things that's like boom, overnight, uh nobody no one's employed anymore.

That's not how these things happen. Even rapid technological change, that's still not how these things happen because people move slowly. They are the slowest changing part. Um Katie and I over at the Trust Insights Podcast are going to be talking about this uh in the not too distant future, that technology moves real fast. People do not, right?

This organic shell um is pretty much the same as it was 50,000 years ago, right? Uh maybe less hair. But there's less uh uh there's less change here than there is in in the the large language model world. These creatures, humans, they don't change very fast. So it's a good question, and there's a lot to keep your eye on.

But if you're agile and you're flexible and you're a lifelong learner, you're gonna be just fine. You're gonna be more than fine. Thanks for the question, and thanks for tuning in. Talk to you next time. If you like this video, go ahead and hit that subscribe button.