I think in the near future, AI isn't going to produce very high-quality art just from a prompt. Rather, it's going to help creative writers by making the research process easier. For example, in this piece of mine, I needed to do research on John Milton and Dante, and that process was greatly facilitated by AI. Also, I occasionally used AI for idea generation: the "Belching Opera" for the Bouville Shenanigans was suggested by AI. But I wrote every word of the piece, and that hard work, hopefully, gives it originality and authenticity.
In the more distant future, who knows? There may come a time when AI can write at the level of a first-rate author. I don't think that that will cause us to lose our humanity. I hope we will come to use AI as an enhancement of our humanity, just as we use other technologies.
I think you are appropriately alarmist. I also think we're headed for an inevitable plunge in the discourse as the enthusiasm of early adopters like Mollick gets replaced by the actual experience of skeptical knowledge workers who are asked to actually do something useful with these tools. There are plenty of people still chirping excitedly at the idea of talking to a talking computer and so far, the disaster unfolding around the LA Unified School District's AI chatbot is not getting much news coverage, but it feels like maybe the cycle will be turning soon. Hyping technology works best at the demo stage, before rank and file users actually get their hands on it.
Usted lo ha descrito perfectamente, la IA es una "herramienta",y, asi como cuando necesitamos clavar un clavo , usamos un martillo, abrir un hueco y utilizamos una pala. Las herramientas ayudan al hombre y para eso han sido creadas. Seria poco inteligente crear , para que esa "creacion" nos reemplacen . He leido sobre los trabajos o profesiones que la IA va a reemplazar, y que el tiempo que "nos sobre" al usar la IA lo podemos emplear en cosas mas productivas. Mi opinion es que no se estan considerando todas las variables, y se estan gestando nuevos problemas, para los cuales no se han diseñado las estrategias de solucion. En este momento, los creadores de IA se enfocan en que trabajos o actividades se emplearan la IA, pero , se ha pensado que hacer con las personas cuyos trabajos van a ser reemplazados por la IA, o estamos esperando que se presente la crisis de desempleados y otros problemas ?
I haven't heard anything about the LA schools chatbot, which really speaks to how it's not getting a lot of attention given how much attention I think I'm paying. I tend to agree that as this stuff is prematurely rolled out and the hype meets reality some rubber will meet the road and we'll find out that things are more complicated, but we're in for a lot of opportunity costs between now and then.
A journalist I talked to after I posted my take told me he thought it was most likely a summer slowdown and not a lack of interest. He was right. The NYT and the LA Times have stories up and I expect other educational journalism organizations will be following with more details based on sources in the LAUSD and the furloughed AllHere employees.
When I was a freelance web developer in the optimistic early days of the internet, people would often contact me with their ideas. Very precious ideas; ideas they were so fearful of having "stolen" that they would communicate them to me in weird obfuscated and anonymized ways, trying to get me to quote on the work without revealing its actual purpose.
Over time, I realized all these ideas were variations of the same few mundane things: searchable indexes, some sort of control panel that merged together some common business tasks, things like that. And these people rarely had anything beyond the original idea; now they wanted someone else to somehow magically make it work. But then, as I asked all the necessary follow-up questions required to actually make that happen, their idea would fall apart in its vagueness, and be exposed as a hollow shell with nothing inside. They knew *what* they wanted to do, but had no idea *how* they wanted to get there.
Modern startup tech-bros are like that, too. They sit in coffee shops and board rooms loud-talking at each other about their ideas, and then they "hire some guys" to actually implement them. But the idea people are very dismissive of those who actually build the systems, the ones up to their necks day after day in all the tedious, complicated minutia of what it takes to convert a vague idea into something real. They think someone can can just slap together some libraries into a functioning system in a few days, that the gaps or contradictions in their idea will somehow just be hand-waved away. There is no respect for the work, for the grind, for the expertise. This bias against actual work is exposed in the way AI is being pushed on us.
Ideas are a dime a dozen. We all have the same ideas. A lot of them are good ("I know! Let's cure cancer!") but they are useless without a detailed plan of execution. Even brilliant AI systems, when fed a vague half-baked idea, will only be able to return a half-baked solution.
Buenos dias, quizas lo que voy a exponer es descabellado por que estoy confundida. Una de las aplicaciones de la IA es la atencion al publico (o estoy confundida ?). el hecho es, que es bastante frustrante establecer comunicacion con estas IA o aplicaciones, porque al parecer trabajan con un catalogo de posibles preguntas, y posibles respuestas. y cuando la pregunta que se hace no esta en el catalogo de preguntas, por supuesto, tampoco esta en el catalogo de respuestas, Esto hace - para mi - que usar estas IA o aplicaciones sea frustrante. Quizas peco de pesimista, pero no creo que vayan a crear una IA que tenga todas las respuestas a todos los problemas habidos y por haber, porque se pretende que la IA piense. creo , que al parecer esa es la intencion de sus creadores, para mi eso es una UTOPIA y un error, y esta bien, que no la puedan crear, porque, eso puede alimentar el apetito de dictadores o psicopatas que se les ocurra apoderarse del mundo, como en la pelicula distopicas.
I have so many ideas for books and films that I have literally dozens of files with thoughts and notes on them sitting on my hard drive. I also have done some work helping other writers do manuscript development work, to help them shape an idea into something that can begin to take shape, but the vast majority of these things never see the light of day because the ideas themselves are very very cheap. I think Mollick's idea is that we can maybe see more of these ideas brought to a tangible form, but the process by which LLMs work doesn't move the original idea forward in any way. It stops everything at the idea stage. Creativity is more than that, IMO.
Yes! Absolutely. I wrote something recently about that, the messiness of the creative human process. One key takeaway I noticed is how often we start out with one idea and then only after the painful process of working through it, we end up doing something much different -- and usually better. If AI was to spew out a polished response to our first idea, we'd miss out on all that. I really don't see how that sort of layered back-and-forth can be effectively automated by even the smartest algorithm.
Buenos dias, opino que en realidad la situacion SI es alarmante, por mucho que se avance en la IA , esta jamas podra igualar, y mucho menos superar al hombre. porque el hombre en su complejidad, es perfectamente imperfecto, y eso esta bien, no nos osrprende que cuando estamos frente a una persona., hombre o mujer que se haya "reconstruido" con cirugias plasticas, implantes y otros afeites artificiales en busca de la perfeccion, su rostro , aunque "perfecto" luce frio y artificial, quitandole ese "no se que" que tiene la belleza natural acompañada de leves o marcadas imperfecciones. Ser hombre es poseer dentro de un cuerpo tres dimensiones: Cuerpo, alma y espiritu, todas estas tambien con sus imperfecciones. con sus detalles. Una creacion escrita, ya sea un ensayo, cuento, poema o historia contiene una mezcla de experiencias. conocimientos, e interpretaciones, particulares, unica. imperfecta . El autor da a luz, es "un parto intelectual", es un pensar y repensar las ideas, borronear,en fin. El resultado de una creacion humana tiene mucho de su creador, no solo sus ideas, su imaginacion, sueños. La IA es vista como una utopia, en este momento recuerdo una parte de un texto en relacion a las investigacion y objetividad y subjetividad . Dicho autor, cuyo nombre no recuerdo, en este momento expuso mas o menos lo siguiente "el hombre en sus opiniones, y en sus creaciones es subjetivo, por que es un sujeto, NO puede ser objetivo, porque no somos objetos" los hombres somos sujetos, la IA siempre sera un objeto, una maquina
I have told you this story before, John, but I am famously prone to repeating myself and anyway it seems relevant here. Years ago, I was on a panel at a book festival with an author whose first novel had been a huge high-concept genre bestseller (it was subsequently made into a big budget movie that didn't do very well, I don't think). Before we went on stage he was lamenting to me that none of his subsequent books came even close to the success of his first, and, in fact, each one did worse than the one before it. He said that he and his agent and his publisher couldn't figure out what was wrong. The publicity and sales departments were at a loss. At the end of our panel we had a Q&A and someone asked for books we had recently read and loved. Each author had several titles they wanted to rave about. And then the bestselling author said, "I don't read fiction. I haven't read a novel since I published (his first novel). I am always writing. I don't have time to read." All the other novelists gasped. And, of course, this was the answer to his problem, but he lacked the self-awareness to understand it. Honestly I think a lot of people who don't read fiction think writing it is easy. I believe the more you read, and the more closely you read, the more you want to write. But you also have a greater understanding of how freaking hard it is to do well.
One of my favorite authors, the Uber prolific Anthony Trollope, said that there is no way of both writing easily and writing well. In other words, as you and various others have said, it's just hard work. And I know this was said before, but I'll happily say it again, that we don't need AI to write novels and paint pictures and play music while we scrub the floors and wash dishes. We need AI to scrub floors and wash dishes while we write novels and paint pictures and play music.
I think in the near future, AI isn't going to produce very high-quality art just from a prompt. Rather, it's going to help creative writers by making the research process easier. For example, in this piece of mine, I needed to do research on John Milton and Dante, and that process was greatly facilitated by AI. Also, I occasionally used AI for idea generation: the "Belching Opera" for the Bouville Shenanigans was suggested by AI. But I wrote every word of the piece, and that hard work, hopefully, gives it originality and authenticity.
https://open.substack.com/pub/jonahogilwy/p/the-blind-chicken-scramble?r=loua&utm_campaign=post&utm_medium=web
In the more distant future, who knows? There may come a time when AI can write at the level of a first-rate author. I don't think that that will cause us to lose our humanity. I hope we will come to use AI as an enhancement of our humanity, just as we use other technologies.
I think you are appropriately alarmist. I also think we're headed for an inevitable plunge in the discourse as the enthusiasm of early adopters like Mollick gets replaced by the actual experience of skeptical knowledge workers who are asked to actually do something useful with these tools. There are plenty of people still chirping excitedly at the idea of talking to a talking computer and so far, the disaster unfolding around the LA Unified School District's AI chatbot is not getting much news coverage, but it feels like maybe the cycle will be turning soon. Hyping technology works best at the demo stage, before rank and file users actually get their hands on it.
Usted lo ha descrito perfectamente, la IA es una "herramienta",y, asi como cuando necesitamos clavar un clavo , usamos un martillo, abrir un hueco y utilizamos una pala. Las herramientas ayudan al hombre y para eso han sido creadas. Seria poco inteligente crear , para que esa "creacion" nos reemplacen . He leido sobre los trabajos o profesiones que la IA va a reemplazar, y que el tiempo que "nos sobre" al usar la IA lo podemos emplear en cosas mas productivas. Mi opinion es que no se estan considerando todas las variables, y se estan gestando nuevos problemas, para los cuales no se han diseñado las estrategias de solucion. En este momento, los creadores de IA se enfocan en que trabajos o actividades se emplearan la IA, pero , se ha pensado que hacer con las personas cuyos trabajos van a ser reemplazados por la IA, o estamos esperando que se presente la crisis de desempleados y otros problemas ?
I haven't heard anything about the LA schools chatbot, which really speaks to how it's not getting a lot of attention given how much attention I think I'm paying. I tend to agree that as this stuff is prematurely rolled out and the hype meets reality some rubber will meet the road and we'll find out that things are more complicated, but we're in for a lot of opportunity costs between now and then.
The74 broke the story. Here is the link: https://www.the74million.org/article/turmoil-surrounds-las-new-ai-student-chatbot-as-tech-firm-furloughs-staff-just-3-months-after-launch/
Here is my take: https://ailogblog.substack.com/p/the-la-unified-school-districts-chatbot
Other than that, crickets.
A journalist I talked to after I posted my take told me he thought it was most likely a summer slowdown and not a lack of interest. He was right. The NYT and the LA Times have stories up and I expect other educational journalism organizations will be following with more details based on sources in the LAUSD and the furloughed AllHere employees.
When I was a freelance web developer in the optimistic early days of the internet, people would often contact me with their ideas. Very precious ideas; ideas they were so fearful of having "stolen" that they would communicate them to me in weird obfuscated and anonymized ways, trying to get me to quote on the work without revealing its actual purpose.
Over time, I realized all these ideas were variations of the same few mundane things: searchable indexes, some sort of control panel that merged together some common business tasks, things like that. And these people rarely had anything beyond the original idea; now they wanted someone else to somehow magically make it work. But then, as I asked all the necessary follow-up questions required to actually make that happen, their idea would fall apart in its vagueness, and be exposed as a hollow shell with nothing inside. They knew *what* they wanted to do, but had no idea *how* they wanted to get there.
Modern startup tech-bros are like that, too. They sit in coffee shops and board rooms loud-talking at each other about their ideas, and then they "hire some guys" to actually implement them. But the idea people are very dismissive of those who actually build the systems, the ones up to their necks day after day in all the tedious, complicated minutia of what it takes to convert a vague idea into something real. They think someone can can just slap together some libraries into a functioning system in a few days, that the gaps or contradictions in their idea will somehow just be hand-waved away. There is no respect for the work, for the grind, for the expertise. This bias against actual work is exposed in the way AI is being pushed on us.
Ideas are a dime a dozen. We all have the same ideas. A lot of them are good ("I know! Let's cure cancer!") but they are useless without a detailed plan of execution. Even brilliant AI systems, when fed a vague half-baked idea, will only be able to return a half-baked solution.
Buenos dias, quizas lo que voy a exponer es descabellado por que estoy confundida. Una de las aplicaciones de la IA es la atencion al publico (o estoy confundida ?). el hecho es, que es bastante frustrante establecer comunicacion con estas IA o aplicaciones, porque al parecer trabajan con un catalogo de posibles preguntas, y posibles respuestas. y cuando la pregunta que se hace no esta en el catalogo de preguntas, por supuesto, tampoco esta en el catalogo de respuestas, Esto hace - para mi - que usar estas IA o aplicaciones sea frustrante. Quizas peco de pesimista, pero no creo que vayan a crear una IA que tenga todas las respuestas a todos los problemas habidos y por haber, porque se pretende que la IA piense. creo , que al parecer esa es la intencion de sus creadores, para mi eso es una UTOPIA y un error, y esta bien, que no la puedan crear, porque, eso puede alimentar el apetito de dictadores o psicopatas que se les ocurra apoderarse del mundo, como en la pelicula distopicas.
I have so many ideas for books and films that I have literally dozens of files with thoughts and notes on them sitting on my hard drive. I also have done some work helping other writers do manuscript development work, to help them shape an idea into something that can begin to take shape, but the vast majority of these things never see the light of day because the ideas themselves are very very cheap. I think Mollick's idea is that we can maybe see more of these ideas brought to a tangible form, but the process by which LLMs work doesn't move the original idea forward in any way. It stops everything at the idea stage. Creativity is more than that, IMO.
Yes! Absolutely. I wrote something recently about that, the messiness of the creative human process. One key takeaway I noticed is how often we start out with one idea and then only after the painful process of working through it, we end up doing something much different -- and usually better. If AI was to spew out a polished response to our first idea, we'd miss out on all that. I really don't see how that sort of layered back-and-forth can be effectively automated by even the smartest algorithm.
Buenos dias, opino que en realidad la situacion SI es alarmante, por mucho que se avance en la IA , esta jamas podra igualar, y mucho menos superar al hombre. porque el hombre en su complejidad, es perfectamente imperfecto, y eso esta bien, no nos osrprende que cuando estamos frente a una persona., hombre o mujer que se haya "reconstruido" con cirugias plasticas, implantes y otros afeites artificiales en busca de la perfeccion, su rostro , aunque "perfecto" luce frio y artificial, quitandole ese "no se que" que tiene la belleza natural acompañada de leves o marcadas imperfecciones. Ser hombre es poseer dentro de un cuerpo tres dimensiones: Cuerpo, alma y espiritu, todas estas tambien con sus imperfecciones. con sus detalles. Una creacion escrita, ya sea un ensayo, cuento, poema o historia contiene una mezcla de experiencias. conocimientos, e interpretaciones, particulares, unica. imperfecta . El autor da a luz, es "un parto intelectual", es un pensar y repensar las ideas, borronear,en fin. El resultado de una creacion humana tiene mucho de su creador, no solo sus ideas, su imaginacion, sueños. La IA es vista como una utopia, en este momento recuerdo una parte de un texto en relacion a las investigacion y objetividad y subjetividad . Dicho autor, cuyo nombre no recuerdo, en este momento expuso mas o menos lo siguiente "el hombre en sus opiniones, y en sus creaciones es subjetivo, por que es un sujeto, NO puede ser objetivo, porque no somos objetos" los hombres somos sujetos, la IA siempre sera un objeto, una maquina
There's so much to love about this post! It's dense with deep thoughts, reasons to utter an Amen! Thank you for the book recs too. Love it all.
Abdulaziz alghwly
Abdulaziz alghouli
I have told you this story before, John, but I am famously prone to repeating myself and anyway it seems relevant here. Years ago, I was on a panel at a book festival with an author whose first novel had been a huge high-concept genre bestseller (it was subsequently made into a big budget movie that didn't do very well, I don't think). Before we went on stage he was lamenting to me that none of his subsequent books came even close to the success of his first, and, in fact, each one did worse than the one before it. He said that he and his agent and his publisher couldn't figure out what was wrong. The publicity and sales departments were at a loss. At the end of our panel we had a Q&A and someone asked for books we had recently read and loved. Each author had several titles they wanted to rave about. And then the bestselling author said, "I don't read fiction. I haven't read a novel since I published (his first novel). I am always writing. I don't have time to read." All the other novelists gasped. And, of course, this was the answer to his problem, but he lacked the self-awareness to understand it. Honestly I think a lot of people who don't read fiction think writing it is easy. I believe the more you read, and the more closely you read, the more you want to write. But you also have a greater understanding of how freaking hard it is to do well.
The Audacity was a great book!
I also love this: "The fact that writing can be hard is one of the things that makes it meaningful."
One of my favorite authors, the Uber prolific Anthony Trollope, said that there is no way of both writing easily and writing well. In other words, as you and various others have said, it's just hard work. And I know this was said before, but I'll happily say it again, that we don't need AI to write novels and paint pictures and play music while we scrub the floors and wash dishes. We need AI to scrub floors and wash dishes while we write novels and paint pictures and play music.