I've become more optimistic about this topic in the medium term (say 5-15 years out), though I am pessimistic about the next 2-5 years.
I suspect that AI will change the nature of white collar jobs rather than eliminate them. There may eventually even be more demand for white collar work than there is now. That isn't to say that demand for certain kinds of roles and tasks won't dramatically decline, but that those will be balanced out by increases in demand for other kinds of roles and tasks. Many of these roles and tasks may be new and ones we can't even imagine right now.
We've seen this in the past with programming. Many people thought higher level programming languages, and then low/no-code apps, would reduce demand for programmers. But that didn't happen, it just changed what projects programmers worked on and what skills programmers needed. Fewer programmers needed to understand the nuts and bolts of how hardware worked, for example. And in fact, because low-level work was automated, this allowed programmers to spend more time on other kinds of projects.
The thing with AI is it will only replace professionals when AI alone is as good or better than a human working with AI tools. I don't see that happening anytime soon for most jobs.
As I said, though, that doesn't mean there won't be short-term pain. Organizations will slowly restructure to incorporate AI, and this will eliminate many jobs. For people in roles that will get eliminated or have demand reduced due to AI, this will be painful. But eventually, some organizations will figure out ways to productively use humans + AI doing roles and tasks we can't even imagine right now, and we'll find a new equilibrium. We don't have filing clerks or old school secretaries anymore, but eventually those roles evolved into the modern administrative assistant.
This is great advice. I submit that the same could be said for the internet, generally — including and especially the mobile internet.
Someone once said that human beings became cyborgs the moment our ancestors picked up their first tool, because every tool is simply a cybernetic extension of our bodies. As humans have grown more cybernetically enhanced with more sophisticated tools — including tools which, in turn, are designed to direct other tools (as computer operating systems direct the various software programs which use them) — we have become relatively less human, and more machine. The trick is to let our tools help us with the work we do, and not to largely do this work themselves — or, rather, do it with minimal input.
If you've ever noticed someone who seems hopelessly lost without their cell phone — or, perhaps if you've ever felt lost yourself under this circumstance — then you know just what I mean…
It’s been less than three years since ChatGPT was introduced. The progress since then has been nothing short of astonishing if you are using the latest models. The people working in AI companies are using their AI to write somewhere around 25% of their own software code right now, and we should expect that number to keep rising as the models get better and data centers get larger.
The people predicting huge job loss are not basing this on where LLM’s are today, but rather where the technology is predicted to get to in another 2-5 years.
Great piece. I would also argue in favor of care careers - not just healthcare but childcare, early education, social work, ministry, psychotherapy. Many of these are poorly paid, in part because of the devaluing of traditional 'feminine' vocations, but we can change that through better policy.
I also wonder how much the push for remote work helps make it easier for employers to shift to bots instead of workers. If you don't get to know your employees personally and you don't have an in-person culture at a workplace that includes collaboration, brainstorming, etc, you can be way more transactional about the folks you hire.
I tried OpenAI’s codex for a conceptually clear programming task that I’d been procrastinating for a while and got a nice API that seems to work. First time vibe coding and I’ll do it again.
The chatbot is no genius, it requires extensive tests to self-check in its development loop and often says things incorrect and denies making mistakes even while fixing them, but it’s faster than any person in a well planned task with the right input.
"For example, if I can feed a hundred Atlantic columns into a machine and then it spits me out a new one that I can’t tell the difference from, maybe those essays weren’t really all that important or thought provoking."
I'm pretty skeptical this would actually work. The average Atlantic article is much better written than most of what AI produces. Also, AI struggles to have a unique perspective on anything, which is really important for good writing.
I think you're right, though. It's telling that we haven't really seen any economy-wide increases in productivity, or even really anecdotal examples of higher quality or more quickly produced products, due to AI. We hear about how it increases personal productivity, but if it's mostly in low value tasks, then that's not going to be a revolution.
Still, though, I think there are a lot of jobs, especially certain kinds of entry-level white collar jobs, that mostly do this kind of low-value work. So I still think there might be some pain in the coming years.
So, a lot of writers play a game where they'll ask ChatGPT to write a column in their own voice, just to see what comes out of it. I think it definitely does a pretty good job picking up on your specific stylistic tics, but it's not gonna churn out a full column.
I've become more optimistic about this topic in the medium term (say 5-15 years out), though I am pessimistic about the next 2-5 years.
I suspect that AI will change the nature of white collar jobs rather than eliminate them. There may eventually even be more demand for white collar work than there is now. That isn't to say that demand for certain kinds of roles and tasks won't dramatically decline, but that those will be balanced out by increases in demand for other kinds of roles and tasks. Many of these roles and tasks may be new and ones we can't even imagine right now.
We've seen this in the past with programming. Many people thought higher level programming languages, and then low/no-code apps, would reduce demand for programmers. But that didn't happen, it just changed what projects programmers worked on and what skills programmers needed. Fewer programmers needed to understand the nuts and bolts of how hardware worked, for example. And in fact, because low-level work was automated, this allowed programmers to spend more time on other kinds of projects.
The thing with AI is it will only replace professionals when AI alone is as good or better than a human working with AI tools. I don't see that happening anytime soon for most jobs.
As I said, though, that doesn't mean there won't be short-term pain. Organizations will slowly restructure to incorporate AI, and this will eliminate many jobs. For people in roles that will get eliminated or have demand reduced due to AI, this will be painful. But eventually, some organizations will figure out ways to productively use humans + AI doing roles and tasks we can't even imagine right now, and we'll find a new equilibrium. We don't have filing clerks or old school secretaries anymore, but eventually those roles evolved into the modern administrative assistant.
"…avoid letting AI de-skill you."
This is great advice. I submit that the same could be said for the internet, generally — including and especially the mobile internet.
Someone once said that human beings became cyborgs the moment our ancestors picked up their first tool, because every tool is simply a cybernetic extension of our bodies. As humans have grown more cybernetically enhanced with more sophisticated tools — including tools which, in turn, are designed to direct other tools (as computer operating systems direct the various software programs which use them) — we have become relatively less human, and more machine. The trick is to let our tools help us with the work we do, and not to largely do this work themselves — or, rather, do it with minimal input.
If you've ever noticed someone who seems hopelessly lost without their cell phone — or, perhaps if you've ever felt lost yourself under this circumstance — then you know just what I mean…
It’s been less than three years since ChatGPT was introduced. The progress since then has been nothing short of astonishing if you are using the latest models. The people working in AI companies are using their AI to write somewhere around 25% of their own software code right now, and we should expect that number to keep rising as the models get better and data centers get larger.
The people predicting huge job loss are not basing this on where LLM’s are today, but rather where the technology is predicted to get to in another 2-5 years.
Great piece. I would also argue in favor of care careers - not just healthcare but childcare, early education, social work, ministry, psychotherapy. Many of these are poorly paid, in part because of the devaluing of traditional 'feminine' vocations, but we can change that through better policy.
I also wonder how much the push for remote work helps make it easier for employers to shift to bots instead of workers. If you don't get to know your employees personally and you don't have an in-person culture at a workplace that includes collaboration, brainstorming, etc, you can be way more transactional about the folks you hire.
I tried OpenAI’s codex for a conceptually clear programming task that I’d been procrastinating for a while and got a nice API that seems to work. First time vibe coding and I’ll do it again.
The chatbot is no genius, it requires extensive tests to self-check in its development loop and often says things incorrect and denies making mistakes even while fixing them, but it’s faster than any person in a well planned task with the right input.
"For example, if I can feed a hundred Atlantic columns into a machine and then it spits me out a new one that I can’t tell the difference from, maybe those essays weren’t really all that important or thought provoking."
I'm pretty skeptical this would actually work. The average Atlantic article is much better written than most of what AI produces. Also, AI struggles to have a unique perspective on anything, which is really important for good writing.
I think you're right, though. It's telling that we haven't really seen any economy-wide increases in productivity, or even really anecdotal examples of higher quality or more quickly produced products, due to AI. We hear about how it increases personal productivity, but if it's mostly in low value tasks, then that's not going to be a revolution.
Still, though, I think there are a lot of jobs, especially certain kinds of entry-level white collar jobs, that mostly do this kind of low-value work. So I still think there might be some pain in the coming years.
So, a lot of writers play a game where they'll ask ChatGPT to write a column in their own voice, just to see what comes out of it. I think it definitely does a pretty good job picking up on your specific stylistic tics, but it's not gonna churn out a full column.
For now
Let me rephrase: It’s not going to churn out a usable column.