Maybe "Story-driven" Model?
Few days ago when I was reading Morgan Housel’s book, The Psychology of Money1, the word “story-driven” in 18th chapter “When You’ll Believe Anything”, absolutely impressed me a lot:
1. The more you want something to be true, the more likely you are to believe a story that overestimates the odds of it being true.
2. Everyone has an incomplete view of the world. But we form a complete narrative to fill the gaps.
I ever heard about the concept of “The Web of Meaning” few years before from Harari’s book, Homo Deus2, so I felt familiar about what Housel says, but, the word, “STORY-DRIVEN”, really interested me. Hard to exactly express what sort of feeling it was in words (maybe inspiring), but I really laughed for a while.
In a second, I thought of the word “data-driven”, both them have the same word-formulation. Machine learning, AI, even all statistic technologies, are generally viewed as data-driven approaches “trained” by data. And although nowadays AI prevails around the world, like large language models ChatGPT or so, many specialists believe that they are not that like human beings, and with no humanity. So, is it possible to have a “story-driven” model? If so, what are differences between “story” and normal “data”, and story is just a word, hence what is more fundamental structure to distinguish “story” and “data”? Can and how the models feel the “story”?
I have no idea about them, but I realize why there exist linguistic researchers start to study AI language models. Maybe there have been many researchers who discuss this kind of models, but probably the word “story-driven” is not that acceptable by most rigorous researches, although it really sounds appealing. But I guess, if pure story-driven model exists, it is likely not “trained” under the single criteria of trying to approaching the optima; moral judgement, life care, or other things about humanism may play a more important role in it.
Interesting anyway.
References