Agreed. At 52, I'm trying to learn how to compose music. I don't learn as quickly as I used to, and even with a musical background, it's a real struggle. I've been tempted many times to give up. At one point I said to myself, "Why bother doing this when AI can do it faster and someday perhaps better than you ever possibly could?" That thought pissed me off so much that I decided I should continue BECAUSE of AI. Because I am human, and I want to make human music. Let it make its robot music. Robot music is not the same.
EXACTLY. And you'll enjoy the process, maybe even come to relish the frustrations as you get better at it. No machine gets that part. I used to think this even before AI, with the cyberpunk idea that you could just have a chip shoved in and suddenly be able to speak another language. It would be an empty acquisition.
That always struck me as... off in the Matrix movies. Neo gets an upload and suddenly it's "I know Kung Fu". You might know the shapes of it, and the moves, but do they really *know* it, know the why and not just the what?
Very informative and interesting. Thank you. I don’t really understand most of AI implications for the future but my nearly 4yrs old granddaughter informed me (with an eye roll) that I don’t have to say “please” and “thank you” to Siri because she’s not human. *Sigh
I've been teaching a Humanities course on AI for the past 4 years. They have to create using AI and then analyze it and assess the ethics of their output. It has been amazing because students KNOW what is good writing/art/music/video and what isn't even if they aren't experts. They can tell what is AI and what isn't (there's a lack of humanness to it that they all call out). But really where they always find new ways of seeing AI is in equity and inclusion -- where it fails and where it succeeds. They give me hope for the future.
It’s all gone a bit Blade Runner and Voight-Kampff and Turing Test, hasn’t it. I find the blurry line between machine learning and artificial intelligence interesting. Automation of repetitive tasks or calculations versus AI “people” and “art”. Monks illustrating manuscripts to typesetting and the printing press to typewriters and photocopiers to word processors to computers to…. In my field, we have machine learning for weather forecasting. Some of this is fine, and the forecasts are good. But it worries me, a little, that the fundamental equations and physics are lacking in a lot of the automated stuff, and the connection between physics and climate change is lost. Pattern recognition is fine until the pattern changes. Then you get six-fingered people and art without soul and weather you weren’t expecting. The whole point of art is one human saying “See?!” and another saying “Yes!”.
Pattern recognition works for many tasks that can then be automated. I’m happy for computers to do these, and we all benefit from them doing so. But I don’t want to watch that awful film clip that’s doing the rounds on social media. I don’t want to read a book written by a machine. The humanity is the point.
Interesting stuff, as always. I have no qualms or quibbles (and certainly no Quibis) on your musings about AI and where it's headed. And your description of early days in the DTP trenches certainly took me back.... I started out designing The Scream Factory magazine in Quark Xpress, then later graduated to Adobe InDesign for book design.
Fellow warriors in the DTP trenches ;-) Yeah, I never made the switch... slugged it out with Quark during the (long) years when it was ****... and still cranking stuff out with it today!
I was there too, at the DTP revolution. That is, I wrote programs to typeset dictionaries. My favorite was writing the code to typesetting a Russian-Icelandic dictionary using LaTeX before it supported eight bit character sets. So I took the jobs from traditional typesetters, then a few years later DTP took the job from me. Apparently, publishers had no problems with copy from Pagemaker looking like stencils produced in a garage in comparison, which is the sort of philistine pig ignorance we've come to expect from people whose brains work about the same way as cash registers do. But i pivoted from typesetting to psychology. Because I wanted to and I could.
But what the new-fangled technology is phasing out, is work for unskilled labor. There'll always be work where thinking is obviously required - work that we don't even know about yet. Smart people adapt. The smartest create new markets and new jobs.
But what I see looming ahead in my business is that the hordes of people who can't learn skilled labor is being funneled into jobs where thinking is required but not obviously so. In my private practice, I see quite a few clients who've had months, sometimes years, of low-cost treatment from psychiatric nurses until they caved in and shelled out dough to receive treatment that actually works.
That's really interesting about how people may wind up faking the thinking jobs... a bit worrying, too. And I think you're about how smart people — at least the ones in the right place at the right time with the right advantages — will generally be able to adapt, and likely prosper. What it'll do to everyone else remains to be seen...
interesting article (and different viewpoint from the usual stuff about AI)
You said "I know for a fact that all my novels got thrown into the LLM hopper for prose."
this is fascinating (and unsettling), have you seen anything that has derived from this ? do you know how they got there? what is your publishers stance on this ?
A while back someone leaked a DB of novels and other material that had gone into one LLM or other... and all my stuff was there. I have seen people try to do a "me" for a joke... the results were terrible :-) Possibly because I've hopped all over the place in terms of genre. I think publishers are just shrugging, to be honest...
I once read a (now ancient) review of my fellow countryman Karl Ove Knausgaard in The Guardian. The reviewer was, I recall, trying to make a point about how a book about the quite ordinary life of a quite unknown Norwegian could create such a sensation.
The gist of the review, I felt, can be summed up as: Knausgaard can write two interesting pages about two sheets of bog roll and I can't and I'm so envious!
So when ChatGPT came, I did a blog post where ChatGPT wrote about two sheets of toilet paper in the styles of Knausgaard, Donald Trump, and Michael Moorcock.
I was impressed - not because this is very good, because it's not, but because it was possible to even try.
Here's a recent attempt at the same, in the style of Michael Marshall Smith. The very first attempt was a bit over the top and felt slightly parodic - I burst out in laughter. Not because it was bad. It worked well as a parody.
Here's after some feedback. I hope we're still on speaking terms.
The Bog Roll
The dripping had been going on for hours. It wasn’t loud, not really, but it had that steady rhythm that wormed its way into your skull. Like a second heartbeat, only one that didn’t belong to you.
The bathroom was dark. I’d left the light off, preferring the vague shapes and outlines you only get from streetlights leaking in through frosted glass. It made things softer somehow. Less real.
The roll was on the floor.
I hadn’t put it there, and I hadn’t seen it when I went to bed. It was lying just to the left of the sink, tilted slightly, one edge resting on a cracked tile. For some reason, that bothered me. It shouldn’t have been tilted. A roll should either lie flat or stand upright. Anything else felt… wrong.
I reached for a cigarette, more out of habit than anything else. The lighter flared in the darkness, briefly illuminating the room. Just long enough to catch the reflection in the mirror.
Something moved.
It could have been nothing—a trick of the light, maybe. But I stayed where I was, cigarette halfway to my lips, staring at the roll as if it would explain itself.
Of course, it didn’t. It was just a roll of paper. Two-ply. White. Innocent.
But I couldn’t shake the feeling that it wasn’t supposed to be there.
As you know, I have big feelings about this, mostly around the copyright infringement. My other major concern is that generative AI has massive climate implications. It's a high price to pay so that people can doodle around. "One query to ChatGPT uses approximately as much electricity as could light one light bulb for about 20 minutes." That's HUGE. Using this techology doesn't have the same implications as when you and I used those old Macs to do DTP. We're at a time in history where catastrophic weather is killing people and destroying entire cities like poor Asheville. IMHO, it's just not worth it. https://www.npr.org/2024/07/12/g-s1-9545/ai-brings-soaring-emissions-for-google-and-microsoft-a-major-contributor-to-climate-change
Oh, definitely true — but in the same way that the Industrial Revolution screwed the planet HARD. I'm just not sure I see that as something that's going to change the behavior of the people making all the money... especially in this political climate, sadly.
Excellent points here, Michael. The human experience of art bouncing off AI which only knows what it knows in terms humans try to specify. Currently, creatives are definitely being pushed to the side - I'm hoping this is simply a somewhat temporary result, and they will be slipping back to being able to be artists in whatever genre, and financially supported. Sooner than later - I hope.
I'm pretty vocal about my feelings against AI in its current form. The way companies are shoehorning it into every single facet of life feels intrusive and abusive. Customer consent doesn't seem to matter to them. Not many people want their toaster to talk to them, or for AI assistants to keep popping up in a piece of software [looking at you Adobe] in front of everything else when I've explicitly instructed them not to.
AI in its current form is only inevitable if we accept it without question. I have such a visceral response every time I hear that AI voiceover (especially the perky woman voice on Instagram). AI images are utterly meaningless to me because they haven't come through the filter of the human spirit, but instead have been generated from a gigantic slush pile of work STOLEN from humans, with all the human quirks and emotions homogenised into something soft and bland, something that gives me the ick on a cellular level. Not to mention the environmental impact, and massive resources being dedicated to it.
I watched 'Perfect Days' the other night and there's no AI on the planet that could ever produce something so quiet and gentle and moving, because it can never understand the tiny and invisible fibres of humanity that connect us all in infinite ways. It could never make the intuitive leaps a person made of flesh and blood and life could.
I don't know what the answer is, but I will continue to be a luddite when it comes to AI. It's an existential threat to artists' livelihood, and I really dislike how art is being undercut and undervalued because of it.
Kinda agree with everything you've said, to be honest. I too am way past the eye-rolling stage of companies trying to wedge AI into everything — which I think shows a bit of desperation on their part, as they wonder what it's actually FOR — and trust me: I'm as furious about having my work stolen as everybody else. And I don't get much consolation from the fact that all the happy coders are going to lose their income soon too, when AI takes THEIR jobs... because the ultra-rich tech kings will just get richer.
I guess my argument, such as it is, is that AI's here, and here to stay — and that creatives have to adapt to that reality, and fast. And perhaps one of the ways of doing that is to make our work even MORE human — like the show you mention — so that audiences don't lose sight of what real creativity and art is like...
Definitely agree that we have to adapt, and I really like your suggestion that the way to do it is to fully lean into our humanness. That feels way more hopeful than when people say 'AI is inevitable and we might as well embrace it'.
Agreed. At 52, I'm trying to learn how to compose music. I don't learn as quickly as I used to, and even with a musical background, it's a real struggle. I've been tempted many times to give up. At one point I said to myself, "Why bother doing this when AI can do it faster and someday perhaps better than you ever possibly could?" That thought pissed me off so much that I decided I should continue BECAUSE of AI. Because I am human, and I want to make human music. Let it make its robot music. Robot music is not the same.
EXACTLY. And you'll enjoy the process, maybe even come to relish the frustrations as you get better at it. No machine gets that part. I used to think this even before AI, with the cyberpunk idea that you could just have a chip shoved in and suddenly be able to speak another language. It would be an empty acquisition.
That always struck me as... off in the Matrix movies. Neo gets an upload and suddenly it's "I know Kung Fu". You might know the shapes of it, and the moves, but do they really *know* it, know the why and not just the what?
Very informative and interesting. Thank you. I don’t really understand most of AI implications for the future but my nearly 4yrs old granddaughter informed me (with an eye roll) that I don’t have to say “please” and “thank you” to Siri because she’s not human. *Sigh
Well tell her that I do say "Thank you" to her. It's not just about Siri, it's good for YOU to show gratitude ;-)
I've been teaching a Humanities course on AI for the past 4 years. They have to create using AI and then analyze it and assess the ethics of their output. It has been amazing because students KNOW what is good writing/art/music/video and what isn't even if they aren't experts. They can tell what is AI and what isn't (there's a lack of humanness to it that they all call out). But really where they always find new ways of seeing AI is in equity and inclusion -- where it fails and where it succeeds. They give me hope for the future.
That's great to hear. And yes, I have faith in us. We will weather this technological storm, and whatever one comes after it.
It’s all gone a bit Blade Runner and Voight-Kampff and Turing Test, hasn’t it. I find the blurry line between machine learning and artificial intelligence interesting. Automation of repetitive tasks or calculations versus AI “people” and “art”. Monks illustrating manuscripts to typesetting and the printing press to typewriters and photocopiers to word processors to computers to…. In my field, we have machine learning for weather forecasting. Some of this is fine, and the forecasts are good. But it worries me, a little, that the fundamental equations and physics are lacking in a lot of the automated stuff, and the connection between physics and climate change is lost. Pattern recognition is fine until the pattern changes. Then you get six-fingered people and art without soul and weather you weren’t expecting. The whole point of art is one human saying “See?!” and another saying “Yes!”.
I like that :-) And *excellent* point about pattern-matching, which is what lies behind so much of this...
Pattern recognition works for many tasks that can then be automated. I’m happy for computers to do these, and we all benefit from them doing so. But I don’t want to watch that awful film clip that’s doing the rounds on social media. I don’t want to read a book written by a machine. The humanity is the point.
And I do suspect it always will be.
Some very interesting points that I look forward to raising in class discussions. This terms Computer Science lessons should be really rather fun!
There's certainly a lot to talk about — you're on the front line!
Interesting stuff, as always. I have no qualms or quibbles (and certainly no Quibis) on your musings about AI and where it's headed. And your description of early days in the DTP trenches certainly took me back.... I started out designing The Scream Factory magazine in Quark Xpress, then later graduated to Adobe InDesign for book design.
Fellow warriors in the DTP trenches ;-) Yeah, I never made the switch... slugged it out with Quark during the (long) years when it was ****... and still cranking stuff out with it today!
I was there too, at the DTP revolution. That is, I wrote programs to typeset dictionaries. My favorite was writing the code to typesetting a Russian-Icelandic dictionary using LaTeX before it supported eight bit character sets. So I took the jobs from traditional typesetters, then a few years later DTP took the job from me. Apparently, publishers had no problems with copy from Pagemaker looking like stencils produced in a garage in comparison, which is the sort of philistine pig ignorance we've come to expect from people whose brains work about the same way as cash registers do. But i pivoted from typesetting to psychology. Because I wanted to and I could.
But what the new-fangled technology is phasing out, is work for unskilled labor. There'll always be work where thinking is obviously required - work that we don't even know about yet. Smart people adapt. The smartest create new markets and new jobs.
But what I see looming ahead in my business is that the hordes of people who can't learn skilled labor is being funneled into jobs where thinking is required but not obviously so. In my private practice, I see quite a few clients who've had months, sometimes years, of low-cost treatment from psychiatric nurses until they caved in and shelled out dough to receive treatment that actually works.
That's really interesting about how people may wind up faking the thinking jobs... a bit worrying, too. And I think you're about how smart people — at least the ones in the right place at the right time with the right advantages — will generally be able to adapt, and likely prosper. What it'll do to everyone else remains to be seen...
interesting article (and different viewpoint from the usual stuff about AI)
You said "I know for a fact that all my novels got thrown into the LLM hopper for prose."
this is fascinating (and unsettling), have you seen anything that has derived from this ? do you know how they got there? what is your publishers stance on this ?
A while back someone leaked a DB of novels and other material that had gone into one LLM or other... and all my stuff was there. I have seen people try to do a "me" for a joke... the results were terrible :-) Possibly because I've hopped all over the place in terms of genre. I think publishers are just shrugging, to be honest...
I once read a (now ancient) review of my fellow countryman Karl Ove Knausgaard in The Guardian. The reviewer was, I recall, trying to make a point about how a book about the quite ordinary life of a quite unknown Norwegian could create such a sensation.
The gist of the review, I felt, can be summed up as: Knausgaard can write two interesting pages about two sheets of bog roll and I can't and I'm so envious!
So when ChatGPT came, I did a blog post where ChatGPT wrote about two sheets of toilet paper in the styles of Knausgaard, Donald Trump, and Michael Moorcock.
https://blog.grendel.no/2023/04/09/dassrulltrilogien-the-bog-roll-trilogy/
I was impressed - not because this is very good, because it's not, but because it was possible to even try.
Here's a recent attempt at the same, in the style of Michael Marshall Smith. The very first attempt was a bit over the top and felt slightly parodic - I burst out in laughter. Not because it was bad. It worked well as a parody.
Here's after some feedback. I hope we're still on speaking terms.
The Bog Roll
The dripping had been going on for hours. It wasn’t loud, not really, but it had that steady rhythm that wormed its way into your skull. Like a second heartbeat, only one that didn’t belong to you.
The bathroom was dark. I’d left the light off, preferring the vague shapes and outlines you only get from streetlights leaking in through frosted glass. It made things softer somehow. Less real.
The roll was on the floor.
I hadn’t put it there, and I hadn’t seen it when I went to bed. It was lying just to the left of the sink, tilted slightly, one edge resting on a cracked tile. For some reason, that bothered me. It shouldn’t have been tilted. A roll should either lie flat or stand upright. Anything else felt… wrong.
I reached for a cigarette, more out of habit than anything else. The lighter flared in the darkness, briefly illuminating the room. Just long enough to catch the reflection in the mirror.
Something moved.
It could have been nothing—a trick of the light, maybe. But I stayed where I was, cigarette halfway to my lips, staring at the roll as if it would explain itself.
Of course, it didn’t. It was just a roll of paper. Two-ply. White. Innocent.
But I couldn’t shake the feeling that it wasn’t supposed to be there.
And that I wasn’t supposed to notice.
As you know, I have big feelings about this, mostly around the copyright infringement. My other major concern is that generative AI has massive climate implications. It's a high price to pay so that people can doodle around. "One query to ChatGPT uses approximately as much electricity as could light one light bulb for about 20 minutes." That's HUGE. Using this techology doesn't have the same implications as when you and I used those old Macs to do DTP. We're at a time in history where catastrophic weather is killing people and destroying entire cities like poor Asheville. IMHO, it's just not worth it. https://www.npr.org/2024/07/12/g-s1-9545/ai-brings-soaring-emissions-for-google-and-microsoft-a-major-contributor-to-climate-change
Oh, definitely true — but in the same way that the Industrial Revolution screwed the planet HARD. I'm just not sure I see that as something that's going to change the behavior of the people making all the money... especially in this political climate, sadly.
Nice one, Cyril!
Thank you!
minor typo: i don’t think that mac ran at 16Hz, or you’d still be waiting for the dropshadow to render.
LOL thank you! Doh...
Excellent points here, Michael. The human experience of art bouncing off AI which only knows what it knows in terms humans try to specify. Currently, creatives are definitely being pushed to the side - I'm hoping this is simply a somewhat temporary result, and they will be slipping back to being able to be artists in whatever genre, and financially supported. Sooner than later - I hope.
Me too — and it's the "being financially supported" piece that's key... though it's always been a struggle!
I'm pretty vocal about my feelings against AI in its current form. The way companies are shoehorning it into every single facet of life feels intrusive and abusive. Customer consent doesn't seem to matter to them. Not many people want their toaster to talk to them, or for AI assistants to keep popping up in a piece of software [looking at you Adobe] in front of everything else when I've explicitly instructed them not to.
AI in its current form is only inevitable if we accept it without question. I have such a visceral response every time I hear that AI voiceover (especially the perky woman voice on Instagram). AI images are utterly meaningless to me because they haven't come through the filter of the human spirit, but instead have been generated from a gigantic slush pile of work STOLEN from humans, with all the human quirks and emotions homogenised into something soft and bland, something that gives me the ick on a cellular level. Not to mention the environmental impact, and massive resources being dedicated to it.
I watched 'Perfect Days' the other night and there's no AI on the planet that could ever produce something so quiet and gentle and moving, because it can never understand the tiny and invisible fibres of humanity that connect us all in infinite ways. It could never make the intuitive leaps a person made of flesh and blood and life could.
I don't know what the answer is, but I will continue to be a luddite when it comes to AI. It's an existential threat to artists' livelihood, and I really dislike how art is being undercut and undervalued because of it.
Kinda agree with everything you've said, to be honest. I too am way past the eye-rolling stage of companies trying to wedge AI into everything — which I think shows a bit of desperation on their part, as they wonder what it's actually FOR — and trust me: I'm as furious about having my work stolen as everybody else. And I don't get much consolation from the fact that all the happy coders are going to lose their income soon too, when AI takes THEIR jobs... because the ultra-rich tech kings will just get richer.
I guess my argument, such as it is, is that AI's here, and here to stay — and that creatives have to adapt to that reality, and fast. And perhaps one of the ways of doing that is to make our work even MORE human — like the show you mention — so that audiences don't lose sight of what real creativity and art is like...
Definitely agree that we have to adapt, and I really like your suggestion that the way to do it is to fully lean into our humanness. That feels way more hopeful than when people say 'AI is inevitable and we might as well embrace it'.