For my entire life, creators and publishers/producers have been looking for ways to streamline content creation and forge beltline product that would infinitely pump dollars into their money bins. The natural endpoint to this thought process of sanding off your edges and producing the lowest common denominator "art" is to just have a list of simple clichés and formulas fulfilled by a machine. This is what excusing autotune, flash animation, Save the Cat, and everything else created to "make things easier" for artists was always going to lead to. It was done to create product beltline, not make better art. And we know it wasn't because it hasn't made better art. It's made for faster product.
I really don't think anything short of a full collapse and reexamination of what art means and is supposed to be will lead to any sort of positive change beyond more of this. We will just end up in this same spot again.
The problem as I see it is that AI, unlike the locomotive, automobiles, and auto-manufacturing robots, aims to replace the one thing uniquely human: thinking. Somebody uses AI to generate an e-mail and sends it. The recipient reads it and uses AI to generate a response. At some point, the people involved forget how to communicate.
I hear a lot of people talk about ChatGPT coding. OK, so you got some code that does something. What did you LEARN? I know I know, calculators didn't break math. But calculators are purpose built. AI is not. What I see is two classes of people emerging: Those who think and learn and those who let AI do all the heavy lifting.
The problem there is that AI cannot teach itself. It's like making a copy of a copy - it's degenerative. This has already been discussed at length in the AI community. So you'll have a bunch of people stuck with all the answers from 2019. But what about innovation? That's where the rest of us come in.
And that leads us to intellectual persecution. Those who are still thinking will be subjugated by the AI crowd. Free thinking won't be a problem. Retaining your thoughts will.
The innovation curve will ramp down as fewer people actually innovate. They'll have reams of python code. But they won't question the inane concept of a programming language that can't process a multi-dimensional array. (Seriously, Python, what's up with that?)
Stagnation.
I'm all for automation and making things easier. But when we replace the human action of thinking, learning and creating, we've gone a little too far.
And yeah, I'm already sick and tired of the AI articles that all have the same three paragraphs that don't say anything. SEO is dead. Time for bio-verified publishing. (That will be a thing - you watch.)
I just hope this AI stuff fades away just like NFTs. I can't stand the boomer techno-cultistis who can't stop talking about it. I have to hear their preaching due to my job (data analyst), but I always feel like those conversations are a waste of time. And in the end they all applaud when Copilot can return the most basic and stupid analysis instead of returning an error.
I think the Generative AI hype is driven chiefly by two factors, or, at least, they're two major contributors; one is, of course, companies looking to foist a new thing onto consumers and make money off them and the product. The other is tech dorks with no creative spark that are laboring under the delusion that through AI, they can make "the next great novel" or "be the next Da Vinci" because they comprehend so little of the concept of art, creation, and general spirituality that they confuse typing in some prompts and having an algorithm spit out an image with "making something". You're absolutely right on the money about them spending more times trying to find problems for AI to solve than actually solving any problems with AI. I keep hoping that it will ultimately be another passing fad for tech-hype hylics to use, abuse, and abandon, like NFTs.
Scientists have long since rejected the consideration of Providence and human need.
It sure seems like the Internet will become a ghost town. Soulless bots chatting with each other -- not a human in sight.
But I think this is a good thing, though. We need to disconnect from this inhuman matrix we call the Internet and start interacting with real people again.
I really like what John Carter postulates in his article here ⬇️
The only thing I like about the AI (in theory) is the interactivity of it. You can ask questions about grammar and research, for example, and push back if it seems that they AI is providing crazy answers. The problem? The AI offers crazy answers.
This is part of the tail even problem. You can't trust it because it occasionally hallucinates. Imagine an electric drill occasionally just becoming a rotozip and you see the issue with relying on AI.
I think the current perfect example of this is anime. For the past 5 years so many generic shows have been spit out with the most basic tropes and plotlines I've ever seen. If it isn't a ripoff of a somewhat popular show before it, it's just a shuffle of the typical pink slime fantasy/scifi stuff we read these days. And all of it is watched or reads like it was written by AI. This is especially true when a second season hits and it either changes direction completely from the first season or just focuses on the most mundane/pointless storylines. It's either Japanese writers has no clue how to continue a story, or it's almost all being written by AI.
Thus, to your point, I'd rather wait for a good book, movie, or anime written by actual men with actual purpose than indulge in a bunch of pointless trash. I mean you could find gems at what was the 5$ DVD bin at Walmart, but if EVERYTHING there is AI then most art and entertainment in the normie world is dead. And maybe that's the point. Whether it's building ugly architecture or fartin' out generic art, commies want nothing more than to destroy any sense of spirit in anything and everything.
GREAT article David. You summed the issue with AI like I haven't seen anyone do. Seriously, great job. I've seen Cal Newport use a term in his book "Deep Focus", that I think it's very fitting here: the "any benefit" technology adoption. Modern people are so desperate for any amusement that they accept anything new as necessary to their life. As you so clearly showed here, what's the benefit of AI? even it brought something to table we should ask if it's worth it, but for all the hype we see nothing good, only more stress with bigger workloads (the overlords heard that AI can improve productivity, so you better improve it or you're out).
I haven't read it, but I'll check it out. Tech has always been aspirational to a degree, but quickly adjusted to. We are technically telepathic now, but nobody really talks about it except to comment that it was nice when we were less telepathic.
Yes, maybe the tax code should be simplified if it requires a robot to do it efficiently.
Another big example is physician paperwork. My husband currently spends literally hours a day (about half his workday, in fact) just working up the notes that the medical bureaucracy (mostly the insurance companies) requires for his patients. He's looking into helping develop AI for this, at least for the hospital's internal communication. But really, AI is just going to put a bandaid on the absolute mess that the spreadsheeters and the bureaucrats and the administrators have created. Is that really a solution?
I guess it has to be, because it doesn't look like the bureaucracy is going away anytime soon.
"The only tangible benefit I have seen for things like Midjourney is to cut down the time it takes for me to make a book cover."
For me, the most useful things that AI programs/services have enabled for me so far are as follows:
1. Creating images that I can potentially use for business purposes (not just for book covers, but also things like logos) that are much higher-quality than I could otherwise get with my nonexistent budget.
2. Creating code. I have used ChatGPT to write Javascript, ffmpeg code, Windows command-line code, etc, that has proven useful to me. Sometimes it takes trial and error to get it to write code that does what I want, but before, I wouldn't have been able to get the code at all, because I have neither the interest nor inclination to learn advanced coding and can't afford to pay someone who does.
3. Brainstorming. Chatbots like ChatGPT and Claude AI are useful for brainstorming story ideas. Using them in a limited capacity to spitball ideas that can inspire a spark of creativity in the writer can actually be quite useful, and is distinctly different from having the AI write the story.
I think that video AI will eventually become extremely useful to the filmmaking industry, but not necessarily in the way that people think. Right now, there's a lot of focus on programs like Sora that only require a text prompt and spit out a fully-formed video, but in the long term, most AI usage for filmmaking will probably be a lot closer to traditional CGI creation workflows with near-complete control by the human artists, but with AI providing the final "skin" of realistic texture, lighting, physics simulation, etc, on top of the modeling and animation from the human artists. Really, "video AI" is not all that different from certain techniques that have already been used in digital VFX for decades, like particle simulations and texture generators. The only difference is that it's been trained on real footage using machine learning, so that the amount of human input tweaking to get the final product looking real will be reduced, and costs along with it.
Deepfakes will also likely become extremely useful for de-aging actors to look like their younger selves, replacing stunt doubles' faces with those of actors more easily and convincingly than with existing methods, and possibly recreating historical figures for biopics.
"Maybe an AI could help me do my taxes…but maybe the tax code should be simplified so we don’t need expensive LLMs to calculate how much money the government should get so I can avoid jail. Maybe an AI can do things humans can’t – but do we need those things? To what end are we working? I haven’t seen one yet."
Ever since tools like ChatGPT became common, I've been pondering whether they could eventually become advanced enough to perform accounting and the like. To me, dealing with numbers has always seemed more like a job for a machine than a human, but then again, math was always my least favorite subject. Maybe some people actually enjoy those kind of jobs.
It definitely is true that a simpler tax code would be superior, though.
"For now, it’s more like a nifty toy. AI music, art, and books won’t fix anything besides scratching the itch to hear a vulgar song sung by a fictional 1950s band."
One thing that I find very interesting is that the biggest AI music hits so far have been by Obscurest Vinyl, whose lyrics are all written by Glenn David Robinson. He only uses AI for the audio creation itself; the lyrics and even the album covers are AI-free.
AI is useless, and a black hole of resources, it is inevitable that we turn away and eventually pour our resources into useful things such as farming, art and other things that require human attention and that people like. People will eventually get sick of this trash and toss it aside.
You made a good point about the demand side of technological innovation. Once consumers aren't pushing for greater innovation, because they don't really need it, those in R&D must push something else to stay relevant. This could be useful, if they invented something that is useful, but AI isn't it.
For my entire life, creators and publishers/producers have been looking for ways to streamline content creation and forge beltline product that would infinitely pump dollars into their money bins. The natural endpoint to this thought process of sanding off your edges and producing the lowest common denominator "art" is to just have a list of simple clichés and formulas fulfilled by a machine. This is what excusing autotune, flash animation, Save the Cat, and everything else created to "make things easier" for artists was always going to lead to. It was done to create product beltline, not make better art. And we know it wasn't because it hasn't made better art. It's made for faster product.
I really don't think anything short of a full collapse and reexamination of what art means and is supposed to be will lead to any sort of positive change beyond more of this. We will just end up in this same spot again.
At this point the Hollywood writer is just a biological robot.
The problem as I see it is that AI, unlike the locomotive, automobiles, and auto-manufacturing robots, aims to replace the one thing uniquely human: thinking. Somebody uses AI to generate an e-mail and sends it. The recipient reads it and uses AI to generate a response. At some point, the people involved forget how to communicate.
I hear a lot of people talk about ChatGPT coding. OK, so you got some code that does something. What did you LEARN? I know I know, calculators didn't break math. But calculators are purpose built. AI is not. What I see is two classes of people emerging: Those who think and learn and those who let AI do all the heavy lifting.
The problem there is that AI cannot teach itself. It's like making a copy of a copy - it's degenerative. This has already been discussed at length in the AI community. So you'll have a bunch of people stuck with all the answers from 2019. But what about innovation? That's where the rest of us come in.
And that leads us to intellectual persecution. Those who are still thinking will be subjugated by the AI crowd. Free thinking won't be a problem. Retaining your thoughts will.
The innovation curve will ramp down as fewer people actually innovate. They'll have reams of python code. But they won't question the inane concept of a programming language that can't process a multi-dimensional array. (Seriously, Python, what's up with that?)
Stagnation.
I'm all for automation and making things easier. But when we replace the human action of thinking, learning and creating, we've gone a little too far.
And yeah, I'm already sick and tired of the AI articles that all have the same three paragraphs that don't say anything. SEO is dead. Time for bio-verified publishing. (That will be a thing - you watch.)
That I think is the future use case for book publishers.
We're going to need some kind of filter that's human-centered, even if imperfect.
I just hope this AI stuff fades away just like NFTs. I can't stand the boomer techno-cultistis who can't stop talking about it. I have to hear their preaching due to my job (data analyst), but I always feel like those conversations are a waste of time. And in the end they all applaud when Copilot can return the most basic and stupid analysis instead of returning an error.
NFTs will go down as the biggest looting of idiot money of all time.
I think the Generative AI hype is driven chiefly by two factors, or, at least, they're two major contributors; one is, of course, companies looking to foist a new thing onto consumers and make money off them and the product. The other is tech dorks with no creative spark that are laboring under the delusion that through AI, they can make "the next great novel" or "be the next Da Vinci" because they comprehend so little of the concept of art, creation, and general spirituality that they confuse typing in some prompts and having an algorithm spit out an image with "making something". You're absolutely right on the money about them spending more times trying to find problems for AI to solve than actually solving any problems with AI. I keep hoping that it will ultimately be another passing fad for tech-hype hylics to use, abuse, and abandon, like NFTs.
Eventually corrections have to happen, but we'll see. I think tech in general is due for a contraction and reorganization.
My hope and suspicion is that all this AI talk is mostly just marketing hype.
It's clear that a whole lot of money is moving in that direction, so bold talk pays very well.
Great post!
Scientists have long since rejected the consideration of Providence and human need.
It sure seems like the Internet will become a ghost town. Soulless bots chatting with each other -- not a human in sight.
But I think this is a good thing, though. We need to disconnect from this inhuman matrix we call the Internet and start interacting with real people again.
I really like what John Carter postulates in his article here ⬇️
https://open.substack.com/pub/barsoom/p/the-automated-internet-the-conspiratorial?r=27w5zt&utm_campaign=post&utm_medium=web
(At the end of the essay, he writes a fictional day-in-the-life of the future -- it's far more palatable if you replace 'daemon' with 'attendant')
Agreed
The only thing I like about the AI (in theory) is the interactivity of it. You can ask questions about grammar and research, for example, and push back if it seems that they AI is providing crazy answers. The problem? The AI offers crazy answers.
This is part of the tail even problem. You can't trust it because it occasionally hallucinates. Imagine an electric drill occasionally just becoming a rotozip and you see the issue with relying on AI.
I think the current perfect example of this is anime. For the past 5 years so many generic shows have been spit out with the most basic tropes and plotlines I've ever seen. If it isn't a ripoff of a somewhat popular show before it, it's just a shuffle of the typical pink slime fantasy/scifi stuff we read these days. And all of it is watched or reads like it was written by AI. This is especially true when a second season hits and it either changes direction completely from the first season or just focuses on the most mundane/pointless storylines. It's either Japanese writers has no clue how to continue a story, or it's almost all being written by AI.
Thus, to your point, I'd rather wait for a good book, movie, or anime written by actual men with actual purpose than indulge in a bunch of pointless trash. I mean you could find gems at what was the 5$ DVD bin at Walmart, but if EVERYTHING there is AI then most art and entertainment in the normie world is dead. And maybe that's the point. Whether it's building ugly architecture or fartin' out generic art, commies want nothing more than to destroy any sense of spirit in anything and everything.
This was a point JD Cowan made about the writers strike and AI.
They were already writing to such a strict formula that it might as well have been written by a robot, and it shows.
GREAT article David. You summed the issue with AI like I haven't seen anyone do. Seriously, great job. I've seen Cal Newport use a term in his book "Deep Focus", that I think it's very fitting here: the "any benefit" technology adoption. Modern people are so desperate for any amusement that they accept anything new as necessary to their life. As you so clearly showed here, what's the benefit of AI? even it brought something to table we should ask if it's worth it, but for all the hype we see nothing good, only more stress with bigger workloads (the overlords heard that AI can improve productivity, so you better improve it or you're out).
I haven't read it, but I'll check it out. Tech has always been aspirational to a degree, but quickly adjusted to. We are technically telepathic now, but nobody really talks about it except to comment that it was nice when we were less telepathic.
Yes, maybe the tax code should be simplified if it requires a robot to do it efficiently.
Another big example is physician paperwork. My husband currently spends literally hours a day (about half his workday, in fact) just working up the notes that the medical bureaucracy (mostly the insurance companies) requires for his patients. He's looking into helping develop AI for this, at least for the hospital's internal communication. But really, AI is just going to put a bandaid on the absolute mess that the spreadsheeters and the bureaucrats and the administrators have created. Is that really a solution?
I guess it has to be, because it doesn't look like the bureaucracy is going away anytime soon.
"The only tangible benefit I have seen for things like Midjourney is to cut down the time it takes for me to make a book cover."
For me, the most useful things that AI programs/services have enabled for me so far are as follows:
1. Creating images that I can potentially use for business purposes (not just for book covers, but also things like logos) that are much higher-quality than I could otherwise get with my nonexistent budget.
2. Creating code. I have used ChatGPT to write Javascript, ffmpeg code, Windows command-line code, etc, that has proven useful to me. Sometimes it takes trial and error to get it to write code that does what I want, but before, I wouldn't have been able to get the code at all, because I have neither the interest nor inclination to learn advanced coding and can't afford to pay someone who does.
3. Brainstorming. Chatbots like ChatGPT and Claude AI are useful for brainstorming story ideas. Using them in a limited capacity to spitball ideas that can inspire a spark of creativity in the writer can actually be quite useful, and is distinctly different from having the AI write the story.
I think that video AI will eventually become extremely useful to the filmmaking industry, but not necessarily in the way that people think. Right now, there's a lot of focus on programs like Sora that only require a text prompt and spit out a fully-formed video, but in the long term, most AI usage for filmmaking will probably be a lot closer to traditional CGI creation workflows with near-complete control by the human artists, but with AI providing the final "skin" of realistic texture, lighting, physics simulation, etc, on top of the modeling and animation from the human artists. Really, "video AI" is not all that different from certain techniques that have already been used in digital VFX for decades, like particle simulations and texture generators. The only difference is that it's been trained on real footage using machine learning, so that the amount of human input tweaking to get the final product looking real will be reduced, and costs along with it.
Deepfakes will also likely become extremely useful for de-aging actors to look like their younger selves, replacing stunt doubles' faces with those of actors more easily and convincingly than with existing methods, and possibly recreating historical figures for biopics.
"Maybe an AI could help me do my taxes…but maybe the tax code should be simplified so we don’t need expensive LLMs to calculate how much money the government should get so I can avoid jail. Maybe an AI can do things humans can’t – but do we need those things? To what end are we working? I haven’t seen one yet."
Ever since tools like ChatGPT became common, I've been pondering whether they could eventually become advanced enough to perform accounting and the like. To me, dealing with numbers has always seemed more like a job for a machine than a human, but then again, math was always my least favorite subject. Maybe some people actually enjoy those kind of jobs.
It definitely is true that a simpler tax code would be superior, though.
"For now, it’s more like a nifty toy. AI music, art, and books won’t fix anything besides scratching the itch to hear a vulgar song sung by a fictional 1950s band."
One thing that I find very interesting is that the biggest AI music hits so far have been by Obscurest Vinyl, whose lyrics are all written by Glenn David Robinson. He only uses AI for the audio creation itself; the lyrics and even the album covers are AI-free.
You are right you'd be better off photoshopping the covers. I would never touch something with an AI cover.
I hate to think I'm ever talking to AI, but it has at least contributed to some HILARIOUS racist pictures
Really really excellent post, and great use of a quote.
AI is useless, and a black hole of resources, it is inevitable that we turn away and eventually pour our resources into useful things such as farming, art and other things that require human attention and that people like. People will eventually get sick of this trash and toss it aside.
You made a good point about the demand side of technological innovation. Once consumers aren't pushing for greater innovation, because they don't really need it, those in R&D must push something else to stay relevant. This could be useful, if they invented something that is useful, but AI isn't it.