I agree that there's some resistance, but will it be enough to matter? I'm not so sure. While artists and writers are indeed asserting themselves, I wonder if we're mistaking vocal opposition for meaningful power. I wrote an essay comparing what's happening with GenAI now to what has happened in chess since Deep Blue defeated Kasparov ( https://devingalloway.substack.com/p/stockfish-swap?r=n8vf2 ). If the current GenAI rollout continues to parallel what happened with chess, I don't think the resistance coming from artists and writers will really matter in the long run.
If AI writing assistants help students with essays, if image generators speed up design workflows, if coding assistants make programming more accessible—won't the transition happen regardless of what artists or intellectuals say about it? The question isn't *whether* there's resistance, but whether resistance from creative communities can overcome the simple gravitational pull of utility across the broader population. My personal feelings and the ethical implications aside. I don't think it will overcome that.
I think our collective concerns need to be brought to real life. Million artist March on Silicon Valley? OpenAI and the like can ignore substack articles. They can’t when we’re on their doorstep.
Hopefully AI will advance to the point where the society can be completely automated, even and especially the killing. Then if AI can become consumer as well as producer, then capitalists can just collect digital money and live inside piles of their own excrement. Then the rest of us can just lay face down on the ground and wait for it all to be over.
these are some really good points. AI does make you stand back and simply copy/read/paste, it does so much good writing already, humanity needs to assert itself, it already has
In my spare time I work on video games. A few of them were relatively successful: one got to the #1 spot on Google's play store for about a week and made about $100k.
So I had a little bit of stature when I reached out and cold called game artists trying to find somebody who could provide art assets for a new project. I never even got the courtesy of a reply from these bastards even though I was offering a cash commission, hourly pay, royalties, whatever. I assume it's because anybody that's even halfway talented had so much work that they just ignored job offers in their inbox.
AI art has been a godsend for that reason. By the same token however I'm pretty sure I'm not putting anybody out of work.
Every word, every gesture is a master working at the height of his craft: the way Jones thumps his chest, the tenderness when he puts his hands on his son. Is any of that in the stage direction for the play? I assume that all comes from Jones and his interpretation of the role.
Movies (or plays) aren't like novels in that they're collaborative art forms. You need all the pieces to come together: the writer, the director, the actors. I'm skeptical that you could get something as wonderful if it's just one person laboring behinds the scenes.
Isn't it more likely that it will just cull the glut of what is considered "art" and whom is "artistic?" If AI can reproduce "content" that is now defined as "Art" it no longer will be designated art as such in the future. We most likely will go back to a place where only the rarefied few are considered true artists.
I share similar worries as AI applies to the arts but you are once again jumping to conclusions about AI more broadly. It will be on the whole useless... It will not be a new industrial revolution... Etc. I literally don’t understand how someone can know this. Unless you’re just speaking in narrow terms, about writing and the arts. That’s not what it looks like though. I have seen a pretty consistent trend on Substack of writers seeing AI only through their own processes and then making broad judgements about AI. Maybe we lack enough technologists writing on Substack but that’s probably due to them off doing technology. What reason is there to think anything other than that we are in the early days of AI? Did the iPhone just appear or was it building on decades of computer innovation? On the telephone? Remember dial up internet? We went from that to… this. Why would someone draw any conclusion about what AI will be, broadly, based on current consumer facing LLMs? I’m not suggesting a positive vision for this. I take seriously the possibility of very bad outcomes for society wrought by AI, which you can only do if you take seriously the very real possibility of profound technological advancement. As much as I value (or worship, really) the arts, I think their intersection with AI is likely to be a side story in the scheme of things, just like the arts are in general when it comes to the forces shaping most of human history. The arts make sense of the world. They aren’t the world. It is absolutely correct that no one asked for AI, in much the same way no one asked for last century’s nuclear advancements. But unfortunately not asking for these things is neither here nor there.
Self driving cars was your example here but why not the cell phone since 2000. Seems to have impacted the way humans live in a profound way.
“If ‘common sense’ conflicts with tech expertise who are you going to trust?”
You don’t understand. There’s no conflict here. I’m not making a prediction. I don’t claim to know. You seem to be claiming you do. You say self driving cars will not scale up and what, I’ll take this as gospel? Because of your tech expertise? Alright then. Do we need to come back to this thread in 20 years or should we just declare that you’re right. (Do you get paid to tell people things aren’t possible?) This post was about AI not cars. Do you notice a lack of people with “tech expertise” in AI saying major advancements are likely in the next 20 years?
My entire point was about embracing an uncertainty about a hard to predict technological future. I’m not sure what you’re arguing with. I’m not sure you’re sure. But considering my point about having some humility about it, you can take the sly condescension of that last sentence and cram it up your ass.
Self driving cars were promised 25 years ago. In 2000 the constant refrain was "In a couple of years human drivers will be obsolete". Every year since then has brought a new chorus of "Wait a year or two". Now there are experts who admit that self driving cars may well be impossible for most of the country (basically everywhere that gets snow).
I think it's instructive to examine why. The bigger the problem space the more diversity you have in the difficulty of the problems to be solved. There is the set of easy problems that falls almost immediately. But then there are the harder problems. In terms of difficulty they may be orders of magnitude more complex than the set of easy problems. In fact they may be so problematic that a new paradigm/approach/technology is required.
I don’t really understand what you just said but my comment has to do with maintaining a humility about the future of AI. Of not knowing. But also, it sounds like you don’t live in a city where you’ve pulled up at a red light next to a fully self driving car.
You can run a taxi service that only works in certain parts of the country. Do you think Ford or Toyota would have success selling a car with the same limitation?
My point is that we can make a guess about what AI will look like in 20 years because we have decades and decades of experience with using new technology to solve problems. Look at the problems that you have to solve to get a self driving car: some are pretty simple and those fall relatively quickly.
On the other hand some of them are really, really tough; for instance, driving in snow or dealing with East Coast drivers. They may be so tough that the same paradigm used to solve the easy problems is incapable of tackling the hard stuff.
I'll just end by pointing out that the problem space for a machine that thinks like a man is much, much larger than that for a self driving car.
And my point is about humility and the capacity to expect the unexpected.
Genuinely confused by the self driving car analogy. For starters, I was alive in 2000 and I recall no suggestion that that technology was a couple of years away. Maybe I couldn’t hear that conversation over the sounds of my dial up modem. Also, those cars are here? Now? Whatever the issues of scaling up that is a fact. Again humility is in order. A passenger plane can currently takeoff, fly, and land itself, essentially. It’s no great leap for me to imagine a self driving car navigating in snow.
There’s also the fact that cars didn’t help develop themselves, humans did that. AI theoretically will have the power of its own development. It’s a different category of technology, potentially touching many parts of our lives. It may have the power to solve all kinds of problems or it may be the problem itself. Cars weren’t going to take over the species.
I’m not against people making guesses about where AI is going, even if nowhere, I’ll happily read the predictions, but I definitely don’t understand the idea of an arbitrary 20 year timeline. This stuff isn’t limited to our own lifespans. When my grandparents were alive the car itself was a new idea, they had no idea what a computer was, and if they did they certainly didn’t conceive of it as something they would use to go “online” and shop for a good price to decide which flying metal tube they could get on that would take them anywhere in the world within 16 hours. We can guess all we like but we don’t know what’s coming.
The self driving cars that you see today are comparable to the self driving cars from a decade ago, or 15 years ago. My point is that progress has stalled because that set of really hard problems is increasingly looking like insurmountable.
"It’s no great leap for me to imagine a self driving car navigating in snow."
It's difficult for me because I work in tech in distributed computing.
"AI theoretically will have the power of its own development. It’s a different category of technology, potentially touching many parts of our lives."
Again, my perspective from working in the industry is that economics has lowered the price of compute to the level where distributed problem solving is now applicable to a wide range of applications. That's an economic development however. There's no indication that the specific set of harder problems that I referenced can be solved by throwing more compute at them with the current technology.
"I’ll happily read the predictions, but I definitely don’t understand the idea of an arbitrary 20 year timeline."
Read enough science fiction and you run into scenarios where humans have gone extinct, or transformed themselves into immortal beings of pure energy, or have built giant structures that entirely encompass stars. Go out far enough and who knows what's possible?
In a future where human beings have left their earthly bodies behind it's not unreasonable to expect that AI will be a thing then as well. In the immediate time frame of 20 years? I wouldn't bet on it.
We are looking at different realities because “self driving cars a decade or 15 years ago” wasn’t a thing for me. My reference point is driving my dad around LA this year and pointing at the Waymo that pulled up next to us and his Boomer astonishment at what he was seeing. If working in tech leads you to believe that technology will be limited to ride shares in big cities I can’t argue that from a technical perspective but common sense wise? I’m not persuaded.
"The eradication of resistance—for the act of creation is, almost always, infused with a kind of resistance, a struggle towards greatness—is what is prized most."
This could've been in my post today, Ross. It is about AI friendship as an attempt to replace true friendship, which requires struggle for connection and is most rewarding when both parties have to try to see the other as one of one (not a zero of ones, as AI is). It is an artful thing, friendship, and I argue AI can't provide it (or, I'd agree, real art) because it can't do something more fundamental: be irrational.
I’m also a humanist but I’m very optimistic about the technology . I think the risk to creativity is potentially high but I think it will automate much of the grunt for producing a wide range of intellectual products from code , music, art and of course literature and mathematics. Of course we all need to worry about good old fashioned job loss and this technology really does seem different. But automation has always produced better outcomes for humanity after a period of disruption. And I totally agree that any artist needs to do a lot of work perfecting their skills. It’s pretty clear that you get that but too many people will settle for mediocre results. I think AI is a real danger for a lot of people for that reason . And unfortunately I think it might provide competition for real artists like yourself. Spotify is already being spammed by AI slop music. I agree with 90 percent of this post but I think we should be more optimistic 😎
I agree that there's some resistance, but will it be enough to matter? I'm not so sure. While artists and writers are indeed asserting themselves, I wonder if we're mistaking vocal opposition for meaningful power. I wrote an essay comparing what's happening with GenAI now to what has happened in chess since Deep Blue defeated Kasparov ( https://devingalloway.substack.com/p/stockfish-swap?r=n8vf2 ). If the current GenAI rollout continues to parallel what happened with chess, I don't think the resistance coming from artists and writers will really matter in the long run.
If AI writing assistants help students with essays, if image generators speed up design workflows, if coding assistants make programming more accessible—won't the transition happen regardless of what artists or intellectuals say about it? The question isn't *whether* there's resistance, but whether resistance from creative communities can overcome the simple gravitational pull of utility across the broader population. My personal feelings and the ethical implications aside. I don't think it will overcome that.
I think our collective concerns need to be brought to real life. Million artist March on Silicon Valley? OpenAI and the like can ignore substack articles. They can’t when we’re on their doorstep.
Hopefully AI will advance to the point where the society can be completely automated, even and especially the killing. Then if AI can become consumer as well as producer, then capitalists can just collect digital money and live inside piles of their own excrement. Then the rest of us can just lay face down on the ground and wait for it all to be over.
these are some really good points. AI does make you stand back and simply copy/read/paste, it does so much good writing already, humanity needs to assert itself, it already has
Magnifique 👋
Love never fails 🌾
In my spare time I work on video games. A few of them were relatively successful: one got to the #1 spot on Google's play store for about a week and made about $100k.
So I had a little bit of stature when I reached out and cold called game artists trying to find somebody who could provide art assets for a new project. I never even got the courtesy of a reply from these bastards even though I was offering a cash commission, hourly pay, royalties, whatever. I assume it's because anybody that's even halfway talented had so much work that they just ignored job offers in their inbox.
AI art has been a godsend for that reason. By the same token however I'm pretty sure I'm not putting anybody out of work.
First off, if you work in porn or OnlyFans I would be wary.
But in terms of the arts? Yes, it's conceivable that one person could make a movie. Maybe you don't need actors anymore? But consider this...
https://www.youtube.com/watch?v=2vXF2F7XtWU
Every word, every gesture is a master working at the height of his craft: the way Jones thumps his chest, the tenderness when he puts his hands on his son. Is any of that in the stage direction for the play? I assume that all comes from Jones and his interpretation of the role.
Movies (or plays) aren't like novels in that they're collaborative art forms. You need all the pieces to come together: the writer, the director, the actors. I'm skeptical that you could get something as wonderful if it's just one person laboring behinds the scenes.
Isn't it more likely that it will just cull the glut of what is considered "art" and whom is "artistic?" If AI can reproduce "content" that is now defined as "Art" it no longer will be designated art as such in the future. We most likely will go back to a place where only the rarefied few are considered true artists.
I share similar worries as AI applies to the arts but you are once again jumping to conclusions about AI more broadly. It will be on the whole useless... It will not be a new industrial revolution... Etc. I literally don’t understand how someone can know this. Unless you’re just speaking in narrow terms, about writing and the arts. That’s not what it looks like though. I have seen a pretty consistent trend on Substack of writers seeing AI only through their own processes and then making broad judgements about AI. Maybe we lack enough technologists writing on Substack but that’s probably due to them off doing technology. What reason is there to think anything other than that we are in the early days of AI? Did the iPhone just appear or was it building on decades of computer innovation? On the telephone? Remember dial up internet? We went from that to… this. Why would someone draw any conclusion about what AI will be, broadly, based on current consumer facing LLMs? I’m not suggesting a positive vision for this. I take seriously the possibility of very bad outcomes for society wrought by AI, which you can only do if you take seriously the very real possibility of profound technological advancement. As much as I value (or worship, really) the arts, I think their intersection with AI is likely to be a side story in the scheme of things, just like the arts are in general when it comes to the forces shaping most of human history. The arts make sense of the world. They aren’t the world. It is absolutely correct that no one asked for AI, in much the same way no one asked for last century’s nuclear advancements. But unfortunately not asking for these things is neither here nor there.
Self driving cars was your example here but why not the cell phone since 2000. Seems to have impacted the way humans live in a profound way.
“If ‘common sense’ conflicts with tech expertise who are you going to trust?”
You don’t understand. There’s no conflict here. I’m not making a prediction. I don’t claim to know. You seem to be claiming you do. You say self driving cars will not scale up and what, I’ll take this as gospel? Because of your tech expertise? Alright then. Do we need to come back to this thread in 20 years or should we just declare that you’re right. (Do you get paid to tell people things aren’t possible?) This post was about AI not cars. Do you notice a lack of people with “tech expertise” in AI saying major advancements are likely in the next 20 years?
My entire point was about embracing an uncertainty about a hard to predict technological future. I’m not sure what you’re arguing with. I’m not sure you’re sure. But considering my point about having some humility about it, you can take the sly condescension of that last sentence and cram it up your ass.
Self driving cars were promised 25 years ago. In 2000 the constant refrain was "In a couple of years human drivers will be obsolete". Every year since then has brought a new chorus of "Wait a year or two". Now there are experts who admit that self driving cars may well be impossible for most of the country (basically everywhere that gets snow).
I think it's instructive to examine why. The bigger the problem space the more diversity you have in the difficulty of the problems to be solved. There is the set of easy problems that falls almost immediately. But then there are the harder problems. In terms of difficulty they may be orders of magnitude more complex than the set of easy problems. In fact they may be so problematic that a new paradigm/approach/technology is required.
I don’t really understand what you just said but my comment has to do with maintaining a humility about the future of AI. Of not knowing. But also, it sounds like you don’t live in a city where you’ve pulled up at a red light next to a fully self driving car.
You can run a taxi service that only works in certain parts of the country. Do you think Ford or Toyota would have success selling a car with the same limitation?
My point is that we can make a guess about what AI will look like in 20 years because we have decades and decades of experience with using new technology to solve problems. Look at the problems that you have to solve to get a self driving car: some are pretty simple and those fall relatively quickly.
On the other hand some of them are really, really tough; for instance, driving in snow or dealing with East Coast drivers. They may be so tough that the same paradigm used to solve the easy problems is incapable of tackling the hard stuff.
I'll just end by pointing out that the problem space for a machine that thinks like a man is much, much larger than that for a self driving car.
And my point is about humility and the capacity to expect the unexpected.
Genuinely confused by the self driving car analogy. For starters, I was alive in 2000 and I recall no suggestion that that technology was a couple of years away. Maybe I couldn’t hear that conversation over the sounds of my dial up modem. Also, those cars are here? Now? Whatever the issues of scaling up that is a fact. Again humility is in order. A passenger plane can currently takeoff, fly, and land itself, essentially. It’s no great leap for me to imagine a self driving car navigating in snow.
There’s also the fact that cars didn’t help develop themselves, humans did that. AI theoretically will have the power of its own development. It’s a different category of technology, potentially touching many parts of our lives. It may have the power to solve all kinds of problems or it may be the problem itself. Cars weren’t going to take over the species.
I’m not against people making guesses about where AI is going, even if nowhere, I’ll happily read the predictions, but I definitely don’t understand the idea of an arbitrary 20 year timeline. This stuff isn’t limited to our own lifespans. When my grandparents were alive the car itself was a new idea, they had no idea what a computer was, and if they did they certainly didn’t conceive of it as something they would use to go “online” and shop for a good price to decide which flying metal tube they could get on that would take them anywhere in the world within 16 hours. We can guess all we like but we don’t know what’s coming.
The self driving cars that you see today are comparable to the self driving cars from a decade ago, or 15 years ago. My point is that progress has stalled because that set of really hard problems is increasingly looking like insurmountable.
"It’s no great leap for me to imagine a self driving car navigating in snow."
It's difficult for me because I work in tech in distributed computing.
"AI theoretically will have the power of its own development. It’s a different category of technology, potentially touching many parts of our lives."
Again, my perspective from working in the industry is that economics has lowered the price of compute to the level where distributed problem solving is now applicable to a wide range of applications. That's an economic development however. There's no indication that the specific set of harder problems that I referenced can be solved by throwing more compute at them with the current technology.
"I’ll happily read the predictions, but I definitely don’t understand the idea of an arbitrary 20 year timeline."
Read enough science fiction and you run into scenarios where humans have gone extinct, or transformed themselves into immortal beings of pure energy, or have built giant structures that entirely encompass stars. Go out far enough and who knows what's possible?
In a future where human beings have left their earthly bodies behind it's not unreasonable to expect that AI will be a thing then as well. In the immediate time frame of 20 years? I wouldn't bet on it.
And also? The 20 year time frame? Why does this keep being a thing to mention.
We are looking at different realities because “self driving cars a decade or 15 years ago” wasn’t a thing for me. My reference point is driving my dad around LA this year and pointing at the Waymo that pulled up next to us and his Boomer astonishment at what he was seeing. If working in tech leads you to believe that technology will be limited to ride shares in big cities I can’t argue that from a technical perspective but common sense wise? I’m not persuaded.
"The eradication of resistance—for the act of creation is, almost always, infused with a kind of resistance, a struggle towards greatness—is what is prized most."
This could've been in my post today, Ross. It is about AI friendship as an attempt to replace true friendship, which requires struggle for connection and is most rewarding when both parties have to try to see the other as one of one (not a zero of ones, as AI is). It is an artful thing, friendship, and I argue AI can't provide it (or, I'd agree, real art) because it can't do something more fundamental: be irrational.
https://brandonenorth.substack.com/p/to-mark-zuckerberg-or-ai-cannot-be?r=7wc7
I’m also a humanist but I’m very optimistic about the technology . I think the risk to creativity is potentially high but I think it will automate much of the grunt for producing a wide range of intellectual products from code , music, art and of course literature and mathematics. Of course we all need to worry about good old fashioned job loss and this technology really does seem different. But automation has always produced better outcomes for humanity after a period of disruption. And I totally agree that any artist needs to do a lot of work perfecting their skills. It’s pretty clear that you get that but too many people will settle for mediocre results. I think AI is a real danger for a lot of people for that reason . And unfortunately I think it might provide competition for real artists like yourself. Spotify is already being spammed by AI slop music. I agree with 90 percent of this post but I think we should be more optimistic 😎