Just out of curiosity: do you start with AI-generated text in your writing? You mentioned that ultimately you rewrite every word, but to rewrite, there has to be something written in the first place.
A couple of random, follow-up thoughts to the above:
It may apply to how we phrase things, which you obviously dodge by rewriting stuff. It may, however, equally apply to creativity. The long-tail fringe ideas won't be suggested by AI, because it is designed to suggest the most likely outcome. So, if we feed ourselves on an AI-suggested starting point, we will necessarily be launching our creative journey from a (relatively) common/average point.
I consciously add "relatively" as I don't think any of your stuff is the generic fluff we're served these days, but still, within our niches, even if narrow, we'll still end up in the relative mainstream.
2. In a comment, someone reported that they run one of my posts through AI, asking, among others, "who else has made a similar point?" Seemingly, I'm not that original, as Eric Ries, Marty Cagan, Teresa Torres, and Dan Kahneman all made very similar claims in the past.
Now, don't get me wrong. I'm flattered by a simple juxtaposition with all of them. Yet, I bet there are hundreds of people who have made similar points, but they're not nearly as prominent, and they're not the usual suspects, so AI won't even suggest them.
And I'd so much prefer to refer to unsung heroes rather than best-known names (even though my continuous referring to Kahneman is an ongoing meme at Lunar—I'm doing as much of that).
Seeding myself with my own random inspiration—links that I stumbled upon—gives me a chance to start away from the most likely point. I can still end up wandering closer to the mean. But if I choose the right direction, the way to the fringes is shorter.
And it's on the fringes where the interesting stuff happens.
Even more so, as we are continuously flooded with bland, generic, AI-generated/not reviewed by a human stuff, which is almost everywhere these days.
The answer is that creativity comes from oblique sources. Sometimes I'm asleep, wake up at three am, and am manically writing in the back of whatever book I'm reading. In the dark, it's a mess, but the ideas persist.
Other times they come from conversations, where, as the fix said, "The wrong word goes in the right ear" and it rolls around for a few days.
The text itself might start, as I said in the vid, from me staring at the camera just lecturing into the void about what I'm thinking about, get transcript, feed it into bot, bot creates mediocre understanding of it and I take this to be like the response I'd get from most people if I just told them that.
THen I force the AI to give me something that is closer to my vision and, in doing so, refine that vision by virtue of the AI's inherent incapability to actually provide a good result.
I've also been experimenting with AI tools that will take chord progressions in things I'm composing and suggest complimentary ones. They are almost always mathematically sound recommendations, but are also boring. Quite often I say, "That's boring, why didn't you do something interesting like Em7? and then I'm like... shit .. Em7. and off I go.
They aren't just the most likely outcome, they are also the most ... arbitrary ... of results sometimes. I think, for me, is to assume that the AI is a functional idiot that its inventors and investors are in love with and that the results will jog my own capabilities, not supplant them.
Just out of curiosity: do you start with AI-generated text in your writing? You mentioned that ultimately you rewrite every word, but to rewrite, there has to be something written in the first place.
A couple of random, follow-up thoughts to the above:
1. Research showing how AI models degrade once they feed upon content generated by themselves: https://www.nature.com/articles/s41586-024-07566-y
It may apply to how we phrase things, which you obviously dodge by rewriting stuff. It may, however, equally apply to creativity. The long-tail fringe ideas won't be suggested by AI, because it is designed to suggest the most likely outcome. So, if we feed ourselves on an AI-suggested starting point, we will necessarily be launching our creative journey from a (relatively) common/average point.
I consciously add "relatively" as I don't think any of your stuff is the generic fluff we're served these days, but still, within our niches, even if narrow, we'll still end up in the relative mainstream.
2. In a comment, someone reported that they run one of my posts through AI, asking, among others, "who else has made a similar point?" Seemingly, I'm not that original, as Eric Ries, Marty Cagan, Teresa Torres, and Dan Kahneman all made very similar claims in the past.
Now, don't get me wrong. I'm flattered by a simple juxtaposition with all of them. Yet, I bet there are hundreds of people who have made similar points, but they're not nearly as prominent, and they're not the usual suspects, so AI won't even suggest them.
And I'd so much prefer to refer to unsung heroes rather than best-known names (even though my continuous referring to Kahneman is an ongoing meme at Lunar—I'm doing as much of that).
Seeding myself with my own random inspiration—links that I stumbled upon—gives me a chance to start away from the most likely point. I can still end up wandering closer to the mean. But if I choose the right direction, the way to the fringes is shorter.
And it's on the fringes where the interesting stuff happens.
Even more so, as we are continuously flooded with bland, generic, AI-generated/not reviewed by a human stuff, which is almost everywhere these days.
All this yes.
The answer is that creativity comes from oblique sources. Sometimes I'm asleep, wake up at three am, and am manically writing in the back of whatever book I'm reading. In the dark, it's a mess, but the ideas persist.
Other times they come from conversations, where, as the fix said, "The wrong word goes in the right ear" and it rolls around for a few days.
The text itself might start, as I said in the vid, from me staring at the camera just lecturing into the void about what I'm thinking about, get transcript, feed it into bot, bot creates mediocre understanding of it and I take this to be like the response I'd get from most people if I just told them that.
THen I force the AI to give me something that is closer to my vision and, in doing so, refine that vision by virtue of the AI's inherent incapability to actually provide a good result.
I've also been experimenting with AI tools that will take chord progressions in things I'm composing and suggest complimentary ones. They are almost always mathematically sound recommendations, but are also boring. Quite often I say, "That's boring, why didn't you do something interesting like Em7? and then I'm like... shit .. Em7. and off I go.
They aren't just the most likely outcome, they are also the most ... arbitrary ... of results sometimes. I think, for me, is to assume that the AI is a functional idiot that its inventors and investors are in love with and that the results will jog my own capabilities, not supplant them.