TL;DR: As part of my "scaling academia" project, I tackle automated idea generation. Finetuned large language model, online machine learning, and semantic search. Poisson point processes!
May 11, 2022·edited May 11, 2022Liked by Jan Hendrik Kirchner
I really like this article, because it covers an important part of the creative process. As a writer, I use many of the techniques described on a daily basis, and I agree that some of that may be automated. However, there is one factor that I'd like to point out that is very important in my experience: A good idea often develops out of a bad one. Almost all great stories have started out as mediocre or even bad ideas nobody cared about.
For example, my youth novel "Boy in a White Room", nominated for the German Youth Literature Award 2018, is about the tenth iteration of the same basic idea ("boy wakes up in a virtual world, not knowing how he got there"). The first iteration I wrote in 2004, before I had published anything. I sent it to a literary agent, who told me that it wasn't good. When I asked for specific feedback, he wrote "we're not a writing school". And of course he was right: The idea wasn't good enough at the time. But it never left me, and I tried again and again to write a novel about it. In 2012, I self-published the fifth iteration or so, set in the world of the computer game Minecraft. It became a bestseller. But it wasn't before "Boy in a White Room" that the idea really came to its full power. The lesson: Creativity isn't really about having ideas, or even implementing them. Creativity is about making good ideas out of bad ones, by iterating over them. Probably that can be automated as well, but it may be more difficult.
Really enjoyed this article and "On Scaling Academia". Seeds of Science (theseedsofscience.org) would be interested in publishing some of your work. Let me know if you are interested, you can email us at info@theseedsofscience.org.
I really like this article, because it covers an important part of the creative process. As a writer, I use many of the techniques described on a daily basis, and I agree that some of that may be automated. However, there is one factor that I'd like to point out that is very important in my experience: A good idea often develops out of a bad one. Almost all great stories have started out as mediocre or even bad ideas nobody cared about.
For example, my youth novel "Boy in a White Room", nominated for the German Youth Literature Award 2018, is about the tenth iteration of the same basic idea ("boy wakes up in a virtual world, not knowing how he got there"). The first iteration I wrote in 2004, before I had published anything. I sent it to a literary agent, who told me that it wasn't good. When I asked for specific feedback, he wrote "we're not a writing school". And of course he was right: The idea wasn't good enough at the time. But it never left me, and I tried again and again to write a novel about it. In 2012, I self-published the fifth iteration or so, set in the world of the computer game Minecraft. It became a bestseller. But it wasn't before "Boy in a White Room" that the idea really came to its full power. The lesson: Creativity isn't really about having ideas, or even implementing them. Creativity is about making good ideas out of bad ones, by iterating over them. Probably that can be automated as well, but it may be more difficult.
Really enjoyed this article and "On Scaling Academia". Seeds of Science (theseedsofscience.org) would be interested in publishing some of your work. Let me know if you are interested, you can email us at info@theseedsofscience.org.
footnote 4 might not be such a bad idea after all: https://keymouse.com/