I will say that you are uniquely in a very very human-centric profession: acquisitions editor, cultural commentator. You are uniquely rewarded for having good instincts and good taste. It is certainly true that your work would be the last kind of job affected by AI directly. But in law, design, customer service, HR, driving, some kinds of education, etc etc… it will be a very different story. It will not all be bad or all be good, but it will change everything.
You: "Have you seen what the latest models can do?” In most cases, the answer is no, I haven’t. ... Pointing out to me the power of the models when I say I don’t want to live in Big Tech’s kingdom is like showing me how sleek and efficient a Glock is when I lament violence."
Matt: "Part of the problem is that most people are using the free version of AI tools. The free version is over a year behind what paying users have access to. Judging AI based on free-tier ChatGPT is like evaluating the state of smartphones by using a flip phone."
You freely admit that you're writing from a position of voluntary ignorance, which is exactly what Matt Shumer warns against doing.
I don't want to live in Big Tech's kingdom either. But what you're doing here is like confidently declaring that the loud purring at the door is a dozen housecats and not a lion. If you can really live without ever opening the door, I guess it doesn't matter. But most of us can't. And we'd like to look through the peephole before we do. To unwrap the metaphor, what would "never opening the door" look like in the modern world? Think the Amish. Even the most countercultural among us aren't the Amish.
What the author calls “the free version” IS THE PRODUCT THAT WAS PREDICTED TWO YEARS AGO TO TRANSFORM OUR ECONOMY.
I am just amazed that more people cannot see what’s happening. “The reason what we said didn’t come true is that you weren’t giving us enough money.” The goalposts just keep moving. It’s more Scientology than science.
18 months ago, I needed a literature survey that crossed quantum mechanics—specifically the Heisenberg Uncertainty Principle—and free will theology. That sort of interdisciplinary task would take a uniquely interdisciplinary grad student days of research. ChatGPT produced it in minutes. It did hallucinate a peer-reviewed paper and once even fabricated a Church Father (on another project) with plausible Cappadocian-sounding quotations. But pressed for citations (as i always do), it caught the error and "apologized".
A year ago, one of my students was building a robot and needed functionality beyond the limits of his programming language. I provided ChatGPT with a single paragraph specification. It returned Python code that was roughly 90% complete: 40-60 hours of dev reduced to minutes. It actually moved a real project forward faster than me and my fellow humans could have.
Neither of these examples proves the maximalist claims made in that essay. But the product your ridiculing here IS ALREADY causing real workflow transformation. My examples are from public models over a year ago. Friends currently in IT describe the same trajectory Matt outlines: "five years ago I wrote code; two years ago I debugged AI-generated code; today I primarily write specs and test what the AI produces." A paralegal I know is already seeing AI reshape job prospects in her field. She is relatively secure because her work is attorney-facing rather than book-facing, yet she admits to using AI for case law review even though she fears how the next generation of paralegals will gain experience. She uses it because it performs well.
When I shared Matt’s article with my late-teenagers today (after reading it on your post, which I also shared with them), I said its timeline may be optimistic but not its direction. I recommended they learn these tools and build their career plans around skills that are human-centered rather than procedural. Even if the most dramatic forecasts fail to materialize, the lower rungs of several professional ladders are already shifting.
Predictions about AI (utopian and dystopian) have certainly been hyperbolic. But that doesn't mean the hype is wholly wrong. As Matt says, the effects are already visible in some fields. They haven't hit yours yet (you are human focused as well) but they will. And sooner than you think.
But still, the anecdote of a working app he gives is from a software developer that's at the forefront of the AI development. I'd rather see a similar article with a perspective of task transformation in some other, less techy field. The translators were in such position about 2 years ago and have already jumped the shark. Are there any lamenting translators articles?
Also, the developers who will be out of work (if not already) will be code monkeys that were already outsourced to India in noughties. But a counterpoint to that is that scandal where an English company offering AI coding no-code agents actually employed Indian developers. So for me... I can't really decide what to bet on. The truth is probably somewhere in between and the loads of money certainly don't help to clear up the situation.
My computer lab teacher in 2007 didn't even bother teaching us how the fax machine works because it is a "dead tech". I use a fax machine every day at work.
Seems like an odd take here, not up to your usual standard. I didn't read that "viral article" as the usual AI boosterism hype, but as a sober "situational awareness" reality check--at least the first half. I would encourage you to re-engage with it from that perspective, and perhaps focus your critiques on the second half, where he adopts the dumb "the best thing you can do now to stay relevant is to really start using AI a lot" framing.
I just can’t take anyone seriously who confidently announces that AI hallucinations are a thing of the past. Just this month attorneys in Kansas were fined because their brief contained hallucinated case law. “That’s the free version, it doesn’t count” is just misdirection, like saying weed isn’t actually harmful, just the bad/cheap kind.
I’m fine to look like a fool in 5 years because I bet against AI ushering in the singularity. But this kind of stuff causes real panic in real people, and every December we’re told “just wait until next year.” I’m fine to be wrong, I just won’t be manipulated.
With Covid, the panic kicked in because people ignored the problem until it was upon them. Then you got all kinds of irrational behavior in both directions. I read this post as saying, "let's not make that same mistake, and start having the serious conversation sooner."
I'm totally with you on not using AI to write. My husband works in tech, and it's interesting to see how he interacts with it vs. how the Substack people feel about it. It doesn't feel humanity destroying to use it to code in the same way that it does to use it to generate thoughts and ideas and writing. It's definitely getting better, but he still needs to do a good bit of fixing and guiding while using it. My guess is that the people that can use it the best will be the people that learned without it (at least with coding).
Speaking as someone who both works in tech and is deeply suspicious of AI: I think 'Is AI as good at certain tasks as people say?', and 'Is AI a good thing, and should we even want it to be good at those tasks?' are two different questions that need to not be conflated. I see that in both the viral article and this response to it. You compare AI to social media, which you say hasn't improved society for the better (I agree), but undoubtedly has changed all sorts of things -- as you yourself describe.
Also too many critiques of AI in the past have majored on 'It's not as good as people say' and pointed to examples of it messing up, which is a problem if AI ever starts cutting those mistakes out. You either have to suddenly pivot to the more crucial issues at the heart of things, like how it is forming us, or start pretending it's not as good at certain things as it is.
Samuel: I think your commentary is provocative and timely for today. As I shudder at the hype and what some are claiming to be so wonderful for humankind as to be like a god, that AI will change the world; it will. But not in the way most writers, editors, poets, thinkers who believe in the power of the creativity of the mind to create and to do it without some machine who is touted to know more than anyone. And its programmed to be that way. I for one will not use it in my writing as I don't believe in the power of dumbing down the mind. Education is lost because kids will use it and will not learn to think or create their own thesis. This is what the powers to be want. Take creativity out of the mind and they'll take your mind and twist it to their algorithm. I'm not a fan nor am I in. I've lived a long time in the real world to see the difference and I hope to continue to do so. I'm bucking the system as long as I can.
Thank God for you Sam! And God is in charge of what happens not AI technologists. I am a print historian/writer thus like you in many ways but the key thing is that we are Christians who know that no invention of fallen humanity can overcome the purposes of the Sovereign Lord whom we know and worship. That is ultimately all that matters.
I feel like there is a middle ground between AI not amounting to much and AI being the Singularity. It clearly makes a difference in some way. For instance, I used Deep Research just last night to create a report on high-paying job opportunities for graduates.
Before AI, I probably would have had to hire a consultant to get that information, or spend hours on my own doing research. Google Gemini gave it to me in 5 minutes.
I think that is genuinely incredible. I've begun using it to play little Dungeons & Dragons-style games, too. They aren't nearly as good as a human GM, but they don't cancel sessions, either.
I don't know where AI is taking us. It probably is not going to become Skynet, but it does seem genuinely disruptive.
Technologies always bring about change, but they don't fundamentally alter what it means to be human and what humans require to create healthy and thriving communities. The miniaturization of the computer and the invention of the internet led to our carrying devices in our pockets that can access the accumulated knowledge of the world and people in far-flung places. These devices have altered our priorities, our habits, the kinds of work many of us do—not for the better, in many cases—but they haven't changed what it means to be creatures created in the image of God, nor have they helped us achieve our deepest desires for connection and love and purpose. More crucially, for all of the benefits these technologies have given to us, they have also introduced a host of harms. Technology is always transformative, but also always flawed and more limited than the visionaries want to believe. AI will not be any different.
I know you were talking about AI. I tend to agree with you. But, as far as having a country left? I’m not so sure it is still the country I grew up in.
I’m trying to come to grips with whether it is the evangelical Christianity that I grew up with. I read the books, I know it is but it sure doesn’t feel like it.
I will say that you are uniquely in a very very human-centric profession: acquisitions editor, cultural commentator. You are uniquely rewarded for having good instincts and good taste. It is certainly true that your work would be the last kind of job affected by AI directly. But in law, design, customer service, HR, driving, some kinds of education, etc etc… it will be a very different story. It will not all be bad or all be good, but it will change everything.
You: "Have you seen what the latest models can do?” In most cases, the answer is no, I haven’t. ... Pointing out to me the power of the models when I say I don’t want to live in Big Tech’s kingdom is like showing me how sleek and efficient a Glock is when I lament violence."
Matt: "Part of the problem is that most people are using the free version of AI tools. The free version is over a year behind what paying users have access to. Judging AI based on free-tier ChatGPT is like evaluating the state of smartphones by using a flip phone."
You freely admit that you're writing from a position of voluntary ignorance, which is exactly what Matt Shumer warns against doing.
I don't want to live in Big Tech's kingdom either. But what you're doing here is like confidently declaring that the loud purring at the door is a dozen housecats and not a lion. If you can really live without ever opening the door, I guess it doesn't matter. But most of us can't. And we'd like to look through the peephole before we do. To unwrap the metaphor, what would "never opening the door" look like in the modern world? Think the Amish. Even the most countercultural among us aren't the Amish.
What the author calls “the free version” IS THE PRODUCT THAT WAS PREDICTED TWO YEARS AGO TO TRANSFORM OUR ECONOMY.
I am just amazed that more people cannot see what’s happening. “The reason what we said didn’t come true is that you weren’t giving us enough money.” The goalposts just keep moving. It’s more Scientology than science.
18 months ago, I needed a literature survey that crossed quantum mechanics—specifically the Heisenberg Uncertainty Principle—and free will theology. That sort of interdisciplinary task would take a uniquely interdisciplinary grad student days of research. ChatGPT produced it in minutes. It did hallucinate a peer-reviewed paper and once even fabricated a Church Father (on another project) with plausible Cappadocian-sounding quotations. But pressed for citations (as i always do), it caught the error and "apologized".
A year ago, one of my students was building a robot and needed functionality beyond the limits of his programming language. I provided ChatGPT with a single paragraph specification. It returned Python code that was roughly 90% complete: 40-60 hours of dev reduced to minutes. It actually moved a real project forward faster than me and my fellow humans could have.
Neither of these examples proves the maximalist claims made in that essay. But the product your ridiculing here IS ALREADY causing real workflow transformation. My examples are from public models over a year ago. Friends currently in IT describe the same trajectory Matt outlines: "five years ago I wrote code; two years ago I debugged AI-generated code; today I primarily write specs and test what the AI produces." A paralegal I know is already seeing AI reshape job prospects in her field. She is relatively secure because her work is attorney-facing rather than book-facing, yet she admits to using AI for case law review even though she fears how the next generation of paralegals will gain experience. She uses it because it performs well.
When I shared Matt’s article with my late-teenagers today (after reading it on your post, which I also shared with them), I said its timeline may be optimistic but not its direction. I recommended they learn these tools and build their career plans around skills that are human-centered rather than procedural. Even if the most dramatic forecasts fail to materialize, the lower rungs of several professional ladders are already shifting.
Predictions about AI (utopian and dystopian) have certainly been hyperbolic. But that doesn't mean the hype is wholly wrong. As Matt says, the effects are already visible in some fields. They haven't hit yours yet (you are human focused as well) but they will. And sooner than you think.
Amen
But still, the anecdote of a working app he gives is from a software developer that's at the forefront of the AI development. I'd rather see a similar article with a perspective of task transformation in some other, less techy field. The translators were in such position about 2 years ago and have already jumped the shark. Are there any lamenting translators articles?
Also, the developers who will be out of work (if not already) will be code monkeys that were already outsourced to India in noughties. But a counterpoint to that is that scandal where an English company offering AI coding no-code agents actually employed Indian developers. So for me... I can't really decide what to bet on. The truth is probably somewhere in between and the loads of money certainly don't help to clear up the situation.
My computer lab teacher in 2007 didn't even bother teaching us how the fax machine works because it is a "dead tech". I use a fax machine every day at work.
Seems like an odd take here, not up to your usual standard. I didn't read that "viral article" as the usual AI boosterism hype, but as a sober "situational awareness" reality check--at least the first half. I would encourage you to re-engage with it from that perspective, and perhaps focus your critiques on the second half, where he adopts the dumb "the best thing you can do now to stay relevant is to really start using AI a lot" framing.
I just can’t take anyone seriously who confidently announces that AI hallucinations are a thing of the past. Just this month attorneys in Kansas were fined because their brief contained hallucinated case law. “That’s the free version, it doesn’t count” is just misdirection, like saying weed isn’t actually harmful, just the bad/cheap kind.
I’m fine to look like a fool in 5 years because I bet against AI ushering in the singularity. But this kind of stuff causes real panic in real people, and every December we’re told “just wait until next year.” I’m fine to be wrong, I just won’t be manipulated.
With Covid, the panic kicked in because people ignored the problem until it was upon them. Then you got all kinds of irrational behavior in both directions. I read this post as saying, "let's not make that same mistake, and start having the serious conversation sooner."
I'm totally with you on not using AI to write. My husband works in tech, and it's interesting to see how he interacts with it vs. how the Substack people feel about it. It doesn't feel humanity destroying to use it to code in the same way that it does to use it to generate thoughts and ideas and writing. It's definitely getting better, but he still needs to do a good bit of fixing and guiding while using it. My guess is that the people that can use it the best will be the people that learned without it (at least with coding).
Speaking as someone who both works in tech and is deeply suspicious of AI: I think 'Is AI as good at certain tasks as people say?', and 'Is AI a good thing, and should we even want it to be good at those tasks?' are two different questions that need to not be conflated. I see that in both the viral article and this response to it. You compare AI to social media, which you say hasn't improved society for the better (I agree), but undoubtedly has changed all sorts of things -- as you yourself describe.
Also too many critiques of AI in the past have majored on 'It's not as good as people say' and pointed to examples of it messing up, which is a problem if AI ever starts cutting those mistakes out. You either have to suddenly pivot to the more crucial issues at the heart of things, like how it is forming us, or start pretending it's not as good at certain things as it is.
Samuel: I think your commentary is provocative and timely for today. As I shudder at the hype and what some are claiming to be so wonderful for humankind as to be like a god, that AI will change the world; it will. But not in the way most writers, editors, poets, thinkers who believe in the power of the creativity of the mind to create and to do it without some machine who is touted to know more than anyone. And its programmed to be that way. I for one will not use it in my writing as I don't believe in the power of dumbing down the mind. Education is lost because kids will use it and will not learn to think or create their own thesis. This is what the powers to be want. Take creativity out of the mind and they'll take your mind and twist it to their algorithm. I'm not a fan nor am I in. I've lived a long time in the real world to see the difference and I hope to continue to do so. I'm bucking the system as long as I can.
Thank God for you Sam! And God is in charge of what happens not AI technologists. I am a print historian/writer thus like you in many ways but the key thing is that we are Christians who know that no invention of fallen humanity can overcome the purposes of the Sovereign Lord whom we know and worship. That is ultimately all that matters.
Absolutely spot-on, Samuel!
As both an engineer and an author myself, I know where algorithms can't leap over the creative chasm. All the can do is re-hash in different clothing.
I have written a short novella called 'Hacker' that might interest you; free download from my site https://bobbook.info And I hope you enjoy it.
I feel like there is a middle ground between AI not amounting to much and AI being the Singularity. It clearly makes a difference in some way. For instance, I used Deep Research just last night to create a report on high-paying job opportunities for graduates.
Before AI, I probably would have had to hire a consultant to get that information, or spend hours on my own doing research. Google Gemini gave it to me in 5 minutes.
I think that is genuinely incredible. I've begun using it to play little Dungeons & Dragons-style games, too. They aren't nearly as good as a human GM, but they don't cancel sessions, either.
I don't know where AI is taking us. It probably is not going to become Skynet, but it does seem genuinely disruptive.
Technologies always bring about change, but they don't fundamentally alter what it means to be human and what humans require to create healthy and thriving communities. The miniaturization of the computer and the invention of the internet led to our carrying devices in our pockets that can access the accumulated knowledge of the world and people in far-flung places. These devices have altered our priorities, our habits, the kinds of work many of us do—not for the better, in many cases—but they haven't changed what it means to be creatures created in the image of God, nor have they helped us achieve our deepest desires for connection and love and purpose. More crucially, for all of the benefits these technologies have given to us, they have also introduced a host of harms. Technology is always transformative, but also always flawed and more limited than the visionaries want to believe. AI will not be any different.
I know you were talking about AI. I tend to agree with you. But, as far as having a country left? I’m not so sure it is still the country I grew up in.
I’m trying to come to grips with whether it is the evangelical Christianity that I grew up with. I read the books, I know it is but it sure doesn’t feel like it.