Good article and I think a very important discussion - some thoughts:
- AI generated content is growing exponentially
- That content is all a function of existing content. While techincally imprecise, this "averaging" of content means that producing mediocre results with generative algorithms is easy and producing striking, compelling results is quite hard
- as the amount of content grows exponentially and models are increasingly trained on AI - generated content, that echo chamber will be harder to escape
- taste, or concepts, or imagination - call it what you want, it will be the human element that must differentiate
i have a hope that such a human element will become broadly desirable as people get tired of remixed, watered down content erupting from generative models
I have that same hope with regard to the human element becoming more desirable.
That's a great point about how new AI content is a function of existing content, so there will be a tendency for content to get more and more ... average ... watered down ... just bad ... I'm not sure what is the right adjective.
Another thought I've had is about how AI probably shouldn't be trusted to create content "into the future." You can probably prompt an LLM to do analysis and then continue to do analysis on the analysis it's already done, thus allowing it to "think into the future." But my guess is, and my experience with using AI chatbots has been, they tend to go off in the "wrong direction" without continual guidance and re-prompting.
So AI has a strong ability to look backward and remix existing content, but the ability to come up with truly new ideas and also to choose what topics to stay focused on—these seem to be abilities that will be uniquely possessed by humans, at least for now.
Agreed. Andreessen said it well: The world is a malleable place. If you know what you want, and you go for it with maximum energy and passion, the world will often reconfigure itself around you much more quickly and easily than you would think.
I read the post on your site. These are great points:
"I believe we are entering the age of the muse. In this age, the expertise required will be less about specific materials or platforms or tools and more about how we shape and steer ideas."
"I can write this, I can record it. Then someone can perfect it and distribute it. We can do this creation→capture→distribution pattern in almost every industry."
"When a human interacts with a creative machine they don’t got to it as they would a scribe or librarian."
"The machine is no longer the blank page it has been forever. Nor is it purely recall. It can make new ideas."
Good article and I think a very important discussion - some thoughts:
- AI generated content is growing exponentially
- That content is all a function of existing content. While techincally imprecise, this "averaging" of content means that producing mediocre results with generative algorithms is easy and producing striking, compelling results is quite hard
- as the amount of content grows exponentially and models are increasingly trained on AI - generated content, that echo chamber will be harder to escape
- taste, or concepts, or imagination - call it what you want, it will be the human element that must differentiate
i have a hope that such a human element will become broadly desirable as people get tired of remixed, watered down content erupting from generative models
I have that same hope with regard to the human element becoming more desirable.
That's a great point about how new AI content is a function of existing content, so there will be a tendency for content to get more and more ... average ... watered down ... just bad ... I'm not sure what is the right adjective.
Another thought I've had is about how AI probably shouldn't be trusted to create content "into the future." You can probably prompt an LLM to do analysis and then continue to do analysis on the analysis it's already done, thus allowing it to "think into the future." But my guess is, and my experience with using AI chatbots has been, they tend to go off in the "wrong direction" without continual guidance and re-prompting.
So AI has a strong ability to look backward and remix existing content, but the ability to come up with truly new ideas and also to choose what topics to stay focused on—these seem to be abilities that will be uniquely possessed by humans, at least for now.
Agreed. Andreessen said it well: The world is a malleable place. If you know what you want, and you go for it with maximum energy and passion, the world will often reconfigure itself around you much more quickly and easily than you would think.
I wrote about this recently, but not yet on Substack - I'll post it here shortly. But early preview: https://nearlyinvisible.com/posts/age-of-the-muse
Thanks for the comment, Dan!
I read the post on your site. These are great points:
"I believe we are entering the age of the muse. In this age, the expertise required will be less about specific materials or platforms or tools and more about how we shape and steer ideas."
"I can write this, I can record it. Then someone can perfect it and distribute it. We can do this creation→capture→distribution pattern in almost every industry."
"When a human interacts with a creative machine they don’t got to it as they would a scribe or librarian."
"The machine is no longer the blank page it has been forever. Nor is it purely recall. It can make new ideas."
And short intro to the idea - https://www.tiktok.com/@nearlyinvisible/video/7313898185309605152