Mind Readings: How AI Will Consume Jobs

In today's episode, let's talk about some of the future of work comments that I've seen recently and had discussed over on LinkedIn. There's sort of two general camps of thought about how artificial intelligence is going to interact with us in work. One is sort of what I would call the Brookings Institute perspective, where they say AI will take tasks, not jobs. And what that means is machines can do individual tasks very narrow scopes very easily, right? Pretty straightforward.

But we as the orchestrators of those tasks are still going to be necessary because it's really challenging to do and glue together a lot of tasks, particularly in jobs that are not purely templated. Now, for example, I used to work at a PR agency, and there was one job, uh the account coordinator job, where there were several people in the agency who just copied and pasted results from Google to spreadsheets all day. That is clearly a job that, yes, you don't even need AI to replace that job. Basic automation would replace that job. So that is a case where, yes, in that specific instance, that job was a task, and it can be replaced.

On the other end, are the wild prognostications of OpenAI's founder, Sam Altman, who went on to say, you know, AI will take 95% of marketing agency jobs. That's a bit of a stretch right now. It's probably pretty unlikely anytime really soon. Because, again, a job is more than just a bundle of individual tasks. There's also a lot of connective stuff that goes in between it.

Are there jobs that are highly templated that uh a model or a network of models could do? Yes. The PR agency example is a good example. I've worked in and with a bunch of agencies, and there are tasks and and even some jobs throughout agencies that, yeah, there's there's room for certainly room for efficiency and process optimization. Will the job overall as a container of tasks go away?

Maybe what is more likely to happen, at least in the near term, is as companies look at their workforces as they're trying to figure out what tasks people do, you will offload more tasks to AI. Copy and paste his worksheet, right? This blog post, analyze this strategy, find anomalies in this spreadsheet, and so on and so forth. What will happen from there is those the time those tasks take will get much much shorter, right? If you say, here's a spreadsheet, identify the anomalies in this budget.

Where that might take a human four hours, eight hours, twelve hours, it might take a machine four seconds, eight seconds, twelve seconds to present its uh findings, and then for a human being to maybe take an hour to validate those findings instead. So if you have, say, 10 full-time employees, each cranking away eight hours a day, 40 hours a week, that's 400 hours of labor. If you cut that by a factor of 10, and you now have 40 hours of work because AI is doing a lot of the individual tasks, it's not that any one job is going to go away entirely, but the amount of labor used will go away, right? You you only need 40 hours of human time to perform that task that used to require 400. And so at companies where profits are more important than than people, you're just gonna throw overboard nine people.

You're gonna say, hey, that one person with the help of AI can do this job. So we don't need the other nine people who are doing that same work because we have a more efficient way of doing it. And even in companies where people are are prioritized equally or ahead of profits, there is still a case to be made that from a sound business practices perspective, it's a good idea to have work be optimal. It's a good idea to get as much work done as quickly as possible while maintaining quality. And if a machine can help you do that, it is irrational to do things the old way just because it's the old way, as opposed to finding the more efficiencies.

The exception is if there is a value proposition and doing it the inefficient way that people are willing to pay for. For example, there is a whole, you know, there's there's entire industries of artisanal handcrafted, this, that, and the other thing. Not made by machines, not made quickly, but people place value on that output and therefore are willing to pay a premium for it. And there are plenty of people who do make a living that way. But that's the likely sort of middle ground between tasks, not jobs, and all jobs going away.

They're too extreme. The middle ground is jobs will go away that will get consolidated down. And the more repetitive your job is, or the the more repetitive your tasks within your job are, the faster those get handed to a machine to do. If you are a paralegal and all you are doing is, you know, case briefings, yeah, that's probably that task is something that you'll be asked to automate. Um if you are a policymaker, and your job is drafting legislation, that is something that, again, a machine's gonna help you do that much faster.

And then will we need as many congressional assistance? Maybe, maybe not. For companies that are more progressive, who put people before or equal to profits, you're going to be looking at a lot of reskilling, a lot of getting people trained in the use of AI so that they can do their jobs better, and you can retain those people, and you can keep paying them and helping them earn a living wage, at the understanding that you're going to have to get those people skilled up pretty quickly. And you may not be hiring more, right? If you have a staff of 10 right now and they do 400 hours of work, but you introduce AI that allows them to do 4,000 hours equivalent of work in that same 400 hours, you're not gonna have to hire people for quite some time.

So that's sort of how AI is likely to consume jobs in the near future. Will that change with things like agent networks? Yes, agent networks, the ability for multiple AIs to work together, kind of like a little virtual office. That is something that can consume more jobs where things are not as templated. As context windows and models get bigger, will that will they consume more complex tasks?

Yes, by nature. Even 18 months out from now, it's is unclear. But big change is coming. So get ready for it. Get ready for it by skilling up by taking the initiative to lead AI initiatives within your company, within your organization, to do a risk assessment to say how at risk is my particular job, how much of it could be offloaded to a machine, and how easily could that be done.

And if it looks easy, give it a try, and then do your best to do your best to be the one who puts your old job out of work and takes on the new job of running the machines that do your old job. That's the show for today. Thanks for tuning in. Talk to you soon. If you enjoyed this video, please hit the like button.

Subscribe to my channel if you haven't already. And if you want to know when new videos are available, hit the bell button to be notified as soon as new content is live.