The dominant approach to artificial intelligence, AI, not only by such institutions as OpenAI, Google and Meta, but seemingly by anyone working in the field, presents issues that have not been addressed to a sufficient extent in the ongoing debates. These issues relate to what AI is in relation to human and other living intelligence, and what AI and the current approaches to AI are lacking in terms of human and living wisdom.
In particularly, I like how you synthesized a lot of the thinkers I've been consuming independently, and framed how machine intelligence fits into this. I'd never heard of Jevons paradox, and it immediately makes so many ideas about ever-increasing consumption come into focus!
Thank you!
I'd love for more people to enjoy this.
To that end, could I offer a few thoughts on what might make this essay even more approachable / enjoyable as a reader? For context, I have a degree in computer science and have done a good deal of work in ML/embedding-based machine intelligence, and I consume a lot of sources mentioned her (eg Daniel Schmachtenberger). I read Substacks like The Great Simplification, which is highly related. So I think I might be your target audience, or very close to it!
This feedback is mostly about structure, not content. Feel free to discard any of it that doesn't feel useful :)
1. Make your main argument clear from the outset: you're missing a handle for readers to use to open the door into your writing.
As a reader, by far the best way to do this for me is in the title. Literally the guaranteed first thing everyone will read in each essay.
Rather than "AI and Living Wisdom" (which tells me little about the argument or the interest of it to me), use a title that gets at the heart of your argument. Admitedly, these are hard to craft and finding a good one takes iteration, feedback, and often several days of trial and error, but here are some rough examples of what I mean:
- "AI is a super smart child. But do we really want to be the parent keeping it under control?"
- "If we keep making AI smart but not wise, we're just going to increasing choas"
- "The current approach to AI: building the scaffolding and saying we have a cathedral"
- "Why are we obsessed with making AI more intelligent, but not wise?"
- Adam does this well (albeit with a comedic bent) in Experimental History (though I often don't agree with his oversimplified arguments regarding science *in* the essays, if I'm being honest!)
I used to think, "Oh, this is just click-bait, that's gross." But great titles aren't click bait, they're "legit bait": legitimizing the value of your argument by offering compelling windows into its depth and relevance.
2. The denser the ideas, the shorter the paragraphs need to be for readers to parse it.
This is especially true on mobile, where the screen size is small and its extra difficult to scan what was said earlier in the page. Reading this essay, despite the depth of ideas, I often just felt like being lost in a wall of text, even when I switched to desktop.
School often taught us that paragraphs should make an atomic argument. This is not that practical: many arguments take a lot of paragraphs to make.
Instead, I like thinking of paragraphs like sets in a larger workout: each asks the reader to make a clearly defined intellectual exertion, and then gives them a break to catch their breath before diving into the next. This essay currently feels a bit like a workout written like: "Do 100 pushups, situps, squats. Then do 50 burpees and 50 back squats. Then..." As a reader, I have to do the additional labor of thinking, "Hmm, ok, so maybe I break this into 10 sets of 10 pushups. Then, because situps are easier for me, maybe 5 sets of 20 situps..." In other words: the way through is ultimately discernible, but it takes a lot of work on the readers part.
In a broader sense, careful paragraph breaks aren't merely for style, they are actually work you do on behalf of the reader, easing their way through your argument.
--
But, beyond those details, I really appreciated this essay, and look forward to reading more from you!
Thank you for the essay! Great minds… :) https://open.substack.com/pub/unexaminedtechnology/p/the-two-is-we-need-to-include-in?r=2xhhg0&utm_medium=ios
Really appreciate this essay, Severin!
In particularly, I like how you synthesized a lot of the thinkers I've been consuming independently, and framed how machine intelligence fits into this. I'd never heard of Jevons paradox, and it immediately makes so many ideas about ever-increasing consumption come into focus!
Thank you!
I'd love for more people to enjoy this.
To that end, could I offer a few thoughts on what might make this essay even more approachable / enjoyable as a reader? For context, I have a degree in computer science and have done a good deal of work in ML/embedding-based machine intelligence, and I consume a lot of sources mentioned her (eg Daniel Schmachtenberger). I read Substacks like The Great Simplification, which is highly related. So I think I might be your target audience, or very close to it!
This feedback is mostly about structure, not content. Feel free to discard any of it that doesn't feel useful :)
1. Make your main argument clear from the outset: you're missing a handle for readers to use to open the door into your writing.
As a reader, by far the best way to do this for me is in the title. Literally the guaranteed first thing everyone will read in each essay.
Rather than "AI and Living Wisdom" (which tells me little about the argument or the interest of it to me), use a title that gets at the heart of your argument. Admitedly, these are hard to craft and finding a good one takes iteration, feedback, and often several days of trial and error, but here are some rough examples of what I mean:
- "AI is a super smart child. But do we really want to be the parent keeping it under control?"
- "If we keep making AI smart but not wise, we're just going to increasing choas"
- "The current approach to AI: building the scaffolding and saying we have a cathedral"
- "Why are we obsessed with making AI more intelligent, but not wise?"
- Real life example from Sundogg substack: "Why Do Rich People In Movies Seem So Fake?" Description: "What we get wrong about class -- and why it matters" https://sundogg.substack.com/p/why-do-rich-people-in-movies-seem
- Adam does this well (albeit with a comedic bent) in Experimental History (though I often don't agree with his oversimplified arguments regarding science *in* the essays, if I'm being honest!)
I used to think, "Oh, this is just click-bait, that's gross." But great titles aren't click bait, they're "legit bait": legitimizing the value of your argument by offering compelling windows into its depth and relevance.
2. The denser the ideas, the shorter the paragraphs need to be for readers to parse it.
This is especially true on mobile, where the screen size is small and its extra difficult to scan what was said earlier in the page. Reading this essay, despite the depth of ideas, I often just felt like being lost in a wall of text, even when I switched to desktop.
School often taught us that paragraphs should make an atomic argument. This is not that practical: many arguments take a lot of paragraphs to make.
Instead, I like thinking of paragraphs like sets in a larger workout: each asks the reader to make a clearly defined intellectual exertion, and then gives them a break to catch their breath before diving into the next. This essay currently feels a bit like a workout written like: "Do 100 pushups, situps, squats. Then do 50 burpees and 50 back squats. Then..." As a reader, I have to do the additional labor of thinking, "Hmm, ok, so maybe I break this into 10 sets of 10 pushups. Then, because situps are easier for me, maybe 5 sets of 20 situps..." In other words: the way through is ultimately discernible, but it takes a lot of work on the readers part.
In a broader sense, careful paragraph breaks aren't merely for style, they are actually work you do on behalf of the reader, easing their way through your argument.
--
But, beyond those details, I really appreciated this essay, and look forward to reading more from you!