Maggie Appleton is a product designer at Ought, an AI research lab that’s working on ways to scale open-ended thinking with machine learning. She’s also an advocate of sharing personal knowledge online. She publishes one of the best digital gardens I’ve seen, located at maggieappleton.com. In this conversation, we discuss the what, why, and how of digital gardening.

Show notes

Show notes include Amazon affiliate links. We get a small commission for purchases made through these links.

If you're enjoying the show, please rate or review us in Apple's podcast directory:
https://podcasts.apple.com/us/podcast/the-informed-life/id1450117117?itsct=podcast_box&itscg=30200

This episode's transcript was produced by an AI. If you notice any errors, please get in touch.

Transcript

Jorge: Maggie, welcome to the show.

Maggie: Hi. Thanks for having me.

Jorge: I am very excited to have you. As I was telling you before we started recording, I’ve been following your work for a while and I have been wowed by your output. Your website, frankly, is inspiring in that it … you put out great content. But also it’s very usable and frankly beautiful. So I wanted to dive into it. But before we do, some folks listening in might not have heard of you or be aware of you. How do you introduce yourself?

About Maggie

Maggie: Sure. I’m getting better at getting this into a concise version. So, I’m a product designer at an AI research lab called Ought, where we’re working on ways to scale open-ended reasoning with machine learning. That really means using what is mostly large language models — you know, the famous AI theme of the moment — to build tools for researchers and really people doing lots of intensive scientific research and academics.

So, that work is really fascinating. I think I’m here today and I’m part of this community of people interested in knowledge management because I got well known for writing about digital gardening on the web, writing about personal knowledge management. I was big into what’s kind of called the note-taking space, although note-taking is obviously like the kind of misguided term for writing and thinking and having an active intellectual life and information rich digital life.

So I think those are the relevant bits. That mostly that I’m a designer and I’m a information management nerd, and I’m a big proponent of people publishing personal knowledge on the web.

Jorge: Well, that’s what I was alluding to when I said that I’ve been wowed by your work, because that’s mostly how I know you through your words primarily. Words, but also drawings. Do you make the drawings yourself?

Maggie: Yeah. So before I worked as a product designer, I was an illustrator. Like, early in my career, for the first couple years. So, I got very good at hand drawn illustration. I did that professionally for a while. I did art direction. And then I got more interested in the kind of UX and product design side of things. But I still have a really strong set of skills in illustration that I pull out when I’m writing essays.

I really love visual explanations. I really love the ways that visuals can make otherwise abstract ideas that we have to grapple with in digital design and programming very tangible and very real, and you can really explain things very well by just anchoring them both in, you know, spatial metaphors, graphics symbols… these are all really powerful tools that are definitely underutilized most of the time.

Jorge: Yeah, I was hoping that we would circle back to metaphors because I get the sense that metaphors are central to your work — both text work and visual work. And… well, first of all, I was hoping that you’d tell us a bit more about the role that metaphor plays in your work.

Grounding abstract understanding through metaphors

Maggie: Sure, sure. I’ve been in love with metaphors for a very long time. I’m glad you picked up on that, Because that’s… I don’t know. I remember at 16, I first read Metaphors We Live By, by George Lakoff and Mark Johnson. Which is probably one of my, like, canonical books that defined my identity or something.

I even… like, my mother studied metaphors in her thesis and kind of passed it on to me it was very much inherited thing. And I kind of became obsessed with this understanding that everything we experience in the world is metaphorical to some degree. Almost all that it’s based in embodiment. That’s another one of my huge themes. I just keep going on about how everything relates to metaphors and embodiment; that our experience in the physical world is the basis of all our abstract understanding.

And so, that is obviously incredibly relevant for people like us, working in programming and digital design and knowledge management where we’re trying to grapple with these big abstract ideas and all this kind of like free-floating digital stuff that sometimes doesn’t feel very grounded in a physical reality.

But our understanding of it always is. So this carries through all the illustration work I did when I worked in like programming education, to working on designing digital products: you obviously are very intentionally crafting metaphors and you’re intentionally crafting concrete representations of abstract concepts.

And yeah, through to all the writing I do and all the explanations I do and the illustration, they all have this metaphorical piece to them.

Jorge: And to give folks an idea of what we’re talking about: one of the challenges of an audio-only medium is that when you start talking about illustrations and pictures, people have to imagine. So, to give folks an idea of what we’re talking about here, just as an example, your latest essay, “The Expanding Dark Forest and Generative AI,“ which I’m hoping we’ll get into.

Maggie: Yes.

Yes

Jorge: Features a beautiful illustration of… I’ll try to describe it and do justice to it. But it is like a slice through the layers that underlie a forest. And you can see the tree canopy with the tree trunks and then underneath the kind of sedimentary layers. And you’re giving each of them names: there’s the dark forest and digital gardens and the cozy web, and these are all kind of strata.

Maggie: Mm-hmm.

Jorge: So that’s an example of these kind of visual metaphors, right?

Maggie: Yeah. Because, like, in that essay I do describe in words the concepts I’m getting at. But people… like, everyone always anchors on images. But that’s a good reason for it, right? Because you’ve kind of distilled the essence of what could have been hundreds of words into something that just fits into a single image.

And obviously there’s lots of layers and density and I put like labels. I mix images and words quite a lot in these illustrations. But it means you’re able to condense much more meaning into a smaller space, which obviously works well on social media; you can share it around. It becomes a bit of a meme that can get passed around.

But, yeah, there’s something really magical about the information density of visuals and graphics, which I would argue is based on the fact that humans are deeply embodied in visual creatures before we were linguistic textural creatures. And so it’s kind of pulling on a much richer, kind of higher bandwidth information channel for us.

Jorge: Just reading in your website about your profile, you’ve lived in many different cities around the world. How has the experience of being an embodied consciousness in different cultures influenced your approach to metaphors and languaging?

Taking in a broad perspective

Maggie: Hopefully a lot. It’s one where I guess I’ve never known a time without being… yeah. Being a very embodied human in a very diverse world. I’m British, but I moved overseas when I was six years old to Southeast Asia and I grew up there and then continued traveling throughout my twenties.

So I kind of didn’t settle back down. I returned to London in my late twenties. I was like, “okay, I need to be a stable adult“. But between those years, I was just very… it was like countries all over the place, all kinds of different cities and people and languages and experiences. I think I learned not to take any kind of human culture or human behavior for granted, right?

You just hold very lightly that there is a correct way to do anything in the world, or that there is a correct way to live or there is a correct way to understand. And I ended up studying anthropology based on this kind of fascination. Trying to figure out how do we analyze this in meaningful ways. How do we write this down and capture our understanding of the diversity of human culture in kind of very tangible ways. So that still fascinates me. That’s… yeah, it just gives you a much broader understanding of what’s possible for humans.

Applying anthropology to cultural phenomena

Jorge: How has your background in anthropology influenced the way that you approach this subject?

Maggie: I hope in lots of ways. There’s some of the kind of the expected stuff where like my degree in it gave me a broader literature to pull on, right? We read all the classics of anthropologists and sociologists and it means you have a lot of theories at hand so that when you’re presented with some interesting behavior of like humans on the web or like approaches to information management, you’re like, oh, okay. I can kind of see how like Erving Goffman’s theory of performance of the self might fit in here, right? Or I can see how Clifford Geertz this series of like, how culture develops, fit in here.

The kind of straightforward one is you just sort of have a lot of really clever, thoughtful thinkers that you’ve read the work of, that you can bring into your analysis. So, you are building on the shoulders of giants. You’re not having to do everything yourself like in terms of analysis and theory. I hope it comes in that way.

And the other I think is this lens where you’ve just learned how to look at cultural phenomena and try to analyze it and break it down and treat it as alien, is maybe the best way to put it. It’s like you look at everything and go, “huh! That’s so strange we do that. I’m not going to take that for granted. I’m really going to question why we do it this way. What are the other possible ways we could have done it? What are the ways we could have done it in the past?“

And so, when I look at anything in the programming world or on the web, it’s like, “oh! This is not like a historical inevitability.“ Like, we have landed in this current moment and this current cultural approach to doing this through a series of historical events that unfolded. But there could have been a million other parts we could have gone down, and we shouldn’t shut down those parts. We should keep exploring them. We should keep ourselves open to the fact that they did once exist. And we don’t just kind of look at the web and go, “oh yeah. Obviously it’s this way. It was always going to be this way.“

Jorge: Well, let’s take a lens to one of those possible paths and how it’s diverged. One of the promises of the web is that it makes publishing accessible to a lot people. It’s something that gives pretty much anyone a reach that people have never had before.

Maggie: Mm-hmm.

Digital gardening

Jorge: Around the turn of the century, the way that that manifested was through web blogs, right?

Maggie: Mm-hmm.

Jorge: These primarily chronological publications where people share various aspects of their lives. And that has evolved, or is evolving. Web blogs are not as big as they used to be a while back. And you’ve written about digital gardening, and some folks listening in might not know about this concept. So, I was hoping that you would talk a bit about what digital gardening is, perhaps in contrast to something like blogging.

Maggie: Yeah, yeah, definitely. So, the core description of digital gardening is it essentially is a personal website, just like a blog. But instead of publishing kind of fully finished posts in a chronological order, it instead has more of a wiki-like structure. If you think of like Wikipedia.

Every page isn’t necessarily tied to a certain date. It’s not chronologically stacked on top of one another. And pages are deeply interlinked. So these are… the qualities are, it’s non chronological. All the pages are deeply interlinked. And it’s very imperfect and you grow it slowly over time.

So in traditional blogging, you put up a post and then you’re done, right? And the next day, or the next week, you write a different post and put it up and it’s done. With gardening, you put up a page and you maybe just put up a few bullet points or a few sentences — a few small snippets of something. And it’s not done yet, right?

You can put a status on it, like, “unfinished.“ Or like, you know, “starting to grow.“ And then, you revisit it later when you update your ideas. You’re like, “oh, I’ve changed my mind about this.“ Or like, “I only had a few bullet points before. Now I have like, you know, two paragraphs to say about this topic.“ And you update it. And so, it’s a very different understanding of how you’re going to interact with this personal website.

It’s not like you posting a stream of updates the way that we’ve been trained to do on platforms like Facebook or Twitter or Medium where you are just doing this… well, maybe not Medium, sorry. Tumblr’s one that’s a bit more chronological. Actually, Medium would count too. But yeah, this chronological posting of content — of finished pieces?

Instead, it is this like growing garden of interlinked content where you are just going back and cleaning up and revisiting and adding to bits over time. Which means that you are trying to update your knowledge over time, right? You are not posting a log of things that happened to you or thoughts you had in a stream. You are being like, “okay. I have a bunch of topics I’m trying to deepen my understanding of. And over time, I’ll go in and I will make those topics better. I’ll make them clearer. I will clarify my understanding. I will share information with other people so they can use it.“

It’s very like pro-open source knowledge, pro sharing everything that you know.

Jorge: As a format for publishing, when seen through this lens, blogging feels to me like a more traditional medium in that it’s serialized and periodical, right? Like Charles Dickens would’ve recognized a blog, I think.

Maggie: Yeah, he would’ve! Yeah, he was blogging.

Jorge: Whereas digital gardens are more native to the medium in that they are full on hypertexts, right?

Maggie: Mm-hmm.

Jorge: So that’s one aspect to this. That said, because blogging might be something that we’re more familiar with. And now, people don’t have blogs, but they have Substacks, which is kind of the same idea.

Maggie: Right. It’s the same thing, yeah.

Jorge: And it might be that people are more used to consuming information in that way. It arrives here a little bit at a time. Even if they’re not continuous — if the subjects are not continuous — like, I’m used to this notion of: there’s a new drop from Jorge or whatever, right?

Maggie: Mm-hmm.

Jorge: Whereas hypertext are — can be — kind of harder to relate to in that sometimes, I encounter a hypertext… Wikipedia is a good example. You click on a Wikipedia page and it’s not clear where you go next. Like you go next wherever your attention leads, right? And I get the sense that Wikipedians have invested a lot of effort into making next steps relatively clear and accessible.

But sometimes when I read people’s digital gardens, I’m like, “what do I do with this?“ Like, “where do I go next?“ And I promise there’s a question here! But your digital garden has a structure that I find very accessible. And I was wondering about the relationship between bottom-up and top down structure in a hyper text, from your perspective. Because like I said, I find yours especially… it’s easy to get into.

The challenge of structuring digital gardens

Maggie: Hmm. That’s interesting. That’s a really good question because navigation and structure is one of these design problems in digital gardens I would say we haven’t solved yet. I think there’s a lot of unsolved problems in the kind of design of digital gardens.

Like, we can get into even infrastructure problems later that there aren’t… it’s not easy to build one of these. Mine is built by hand with a bunch of janky code and hodgepodge over the years. Like it is not something I can just hand to someone else and have them spin up their own version of. Which makes me really sad. I wish I was a better developer to like, be able to enable more people to build these. At the moment, we just don’t have great frameworks or systems or principles for doing it.

Navigation… yeah, it’s definitely one of these problems in that the problem you described where you land on a single page, and usually what people do is they put back links somewhere, either on a sidebar or below the post. And these are like pages that link to the page you’re currently on. And that’s really great. That gives you something else to click on next, right? You can go click through and explore. But sometimes you get stuck on a dead end page. There aren’t any links back to it.

And that’s when you need somewhat to be able to jump back out to some sort of global navigation structure, right? Like they need to go to some index page where they can see everything. And if someone has like a thousand notes on a digital garden, you can’t really browse an index page that well. So, that’s where you have to rely on things like filters, searching, you know, our kind of typical design patterns.

We have the searching and browsing information on the web. I wish there were better information architecture patterns. Yeah, I think… again, I’m going to say this is like an unsolved problem at the moment because the chronological stream gives you a very natural order of where to go next, right? You’re just going to the next post. The next post.

And the dream of digital gardens is that actually you should be seeing the most relevant next content every single time you’re on a page, because everything is connected by content type and relationships and themes and tags. But I think that there maybe aren’t good examples of this being done on many digital gardens because we’re missing that sort of infrastructure piece and we’re missing the kind of best design practices

Jorge: When you were talking about anthropology earlier, you mentioned the phrase “performance of the self.“ And I am not versed in anthropology, so I don’t know the text that that refers to. But I’m just going to go off the phrase, okay?

Maggie: Yeah, yeah.

Jorge: So, Andy Matuschak in his digital garden has a note about working with the garage door open.

Maggie: Mm-hmm. I love that one.

Jorge: Right? And people have been thinking like this, in this kind of hypertext format for a while, right? So, Niklas Luhmann with his cards and all that stuff, which… you know, they were not digital, right? Like they were index cards. And the question is, why publish this stuff? Because…

Maggie: hmm. Yeah.

Jorge: … is it different to do it for yourself than to share it with the world?

Why publish this stuff

Maggie: Andy Matuschak is a really, really good example of this. So for people who don’t know, notes.andymatuschak.com is one of my favorite websites. Andy’s a wonderful thinker. He publishes all these small notes and they’re very like, canonical little ideas and none of them are very long. But like every time I happen on one, I feel like I have to sit with it for like an hour because it’s like, “oh, this is such a good insight! This is so rich.“ And part of that is just Andy really has a great mind.

But you know, I aspire to that sort of thing with my garden. Where, when I put up a note… it can be so small, I think is the difference. In our traditional notion of blogging, people have this notion like, “oh, if I’m going to make a blog post, like I have to make a proper post. I can’t do like two sentences, right? I have to do like an actual, I have to give a title that has to be some featured image, you know. I’ve got to like set the tags up.“

It’s like this whole production to post this thing. As publishing to the web has become more commercialized and we’ve all kind of taken on these like big personal brands, it starts to feel very performative, right? Like, there’s all this pressure to put up really impactful posts. Or I think this happens with Substack at the moment, right? People publish things in Substack and think like, “oh God! This post has to be profound and insightful. And I’m going to email it to all these people! I’m doing a push notification!“

That’s, in a way, a very invasive thing to do to other people. It better be worth it! It better be a good quality thing I’m sending to them. Versus digital gardening, is not push based. It’s very, in ways, passive in that you can put up very small things that are totally unpolished and like, might not seem worth publishing. But because it’s so low friction, you are just like, “oh, I’ll just put up a small note about this. It’s not a big deal. Nobody’s going to get push notification.“

In an ideal world, they come across it through something like Google search, right? They’re searching for this exact topic that I’ve written a small note on, and they just find the information they need at the right time. But it’s not like they’re being pushed this out, this big essay.

So, yeah. I’d say it’s low friction, so it allows people just share easily without thinking too much about it or caring too much about this kind of like performance anxiety of the modern web. And it allows people to share much more niche stuff. It’s like you don’t have to pander to the crowd because again, it’s like you’re not pushing a Substack out to a couple hundred people. You are quietly publishing this small thing.

So, you could post the most like niche, tiny, little piece of information about like how you fix your brand of toaster in a certain way. And there’s no cost to it. It’s so easy that then if someone who needs that specific piece of information finds it, you’ve done a great service to them. So it’s… I don’t know. I love the reciprocal sharing of it too. If it were to be private you’re not really sharing everything you know with everyone else. Which I think we all should be doing.

Jorge: Well, I mean, and in your digital garden you have… yeah. They’re not comments. They’re called “Mentions Around The Web,“ which I get the sense that

Maggie: Hmm.

Jorge: you’re pulling from Twitter.

Maggie: Yes. Yeah. That piece is done using web mentions, which is a protocol that allows you to track when other people mention your link. So, if you’re on one of my garden notes, the link to that… I’m kind of scanning the web for when other people mention that. So I scan Twitter. And when I say I scan, that is the web mentions protocol is looking for that specific URL that I’ve given it. And when it finds another website that mentions that url, it will link it in the bottom of mine. So, if you on your blog were to write about one of my pieces and link to it, it would show up at the bottom of mine. So you’re getting this like back links, but across the whole web.

Jorge: Well, the effect of that is that it opens up the note to other voices, right? So there’s a

Maggie: Mm-hmm.

Jorge: bit of feedback involved, but it also kind of invites other points of view, other perspectives. And I want to call out — because this is a podcast where a lot of our focus here is on information architecture and how to structure these things. And I said earlier that your digital garden, I find particularly easy to get into. And I’m going to call it out. I think that part of it is because it is stratified. There are parts of it that feel more note-like and more emergent. And then there are parts of it that feel more produced. Like, I talked about the Expanding Dark Forest essay, and that’s literally tagged as an essay. So, you make a distinction between essays, notes, patterns, library and anti-library.

Maggie: Mm-hmm. Yeah. Yeah.

Jorge: So there is this kind of top-down structure to it, but it’s a top-down structure that accommodates bottom-up emergent thinking, right?

Digital gardens for emergent thinking

Maggie: Yeah, it’s my best solution at the moment. One of my goals in the coming years is to figure out better design solutions for this. So, this is my current one. It’s like my more fleeting, you know, small things I put up as notes. And I also have something called a growth stage on every single one of my posts, which go from seedling to budding to evergreen.

So, seedlings I kind of explicitly say are like, this is tiny, or it’s bad, or I haven’t thought about it. This is just like a fleeting idea I’m going to pop up and I’m going to revisit it later. If it’s somewhere in the middle, it goes to budding and when it’s really like, “okay, I have thought about this. I’ve polished it. I’ve really worked on it,“ I will mark it as an evergreen.

So between the content type, which is like essays and notes and patterns and these growth stages, I have a lot of ways to let readers know how seriously to take content and like signal how much effort I’ve put into them. So everything starts as notes, and then if a note really gets quite big and I’ve polished it and it’s kind of almost an evergreen? I will then turn it into an essay. And essays are much more like: this is my opinion. I’m quite sure of my opinion. I’ve really done the research. Now here’s all my references… that’s more essay shaped. Whereas notes are much more like, oh! Like, here’s a few links. I was thinking about this. I’m not really sure what I think yet.

Jorge: I would imagine that that gives you permission to… what’s that phrase, “strong ideas loosely held,“ right? It gives you permission to…

Maggie: Yeah.

Jorge: So, I really like that. You know, I think I might I might copy that pattern because it’s…

Maggie: Do it. Yeah!

Jorge: I really love this idea of being frank with the reader and saying, “look, you know, this is just an exploration.“ And you are circling back and working on it. All the material that we’re talking about… like one of the distinctions that I think of when I think of keeping notes for yourself versus publishing them in this way, is that in publishing them, they are becoming part of the corpus of the web, which as we know is being mined by the large language models.

Maggie: Yeah.

Jorge: So, I was hoping that we would talk a bit about generative AI and your concept of the dark forest and how this plays into that.

Trust and truth and generative AI

Maggie: Yeah.. I stepped into the AI world about 10 months ago, and it’s been a bit of a jarring experience, I mean, for everyone, right? The last six months have been a bit shocking. My, like, position before language models appeared on the scene was like, we should all publish everything all the time.

Publishing your knowledge to the web both opens you up to have relationships with other people, right? I think I’ve had so many wonderful friendships and collaborations and amazing jobs all come through writing on my website, and writing on Twitter. And I would just, I would say that’s, you know, so invaluable.

There’s nothing I could trade it for. It’s just been the best people. Because it’s like putting out a bat signal for everyone into the same things as you. And they like come running and you’re like, “oh yeah! These are my people!“ I’ve absolutely loved that, and I couldn’t have done that without publishing to the open web and just kind of inviting anyone who wants to talk to me to come chat to me.

And now, we’re facing this moment where language models are scraping the web for all this text and training on it. And we are not quite sure of the repercussions of that yet. But in my essay I’d written, I was mostly worried about how this will affect human relationships on the web.

So, the thing that I really valued from publishing so much. And then trust and truth, I think, are kind of up for debate. Because what happens is that now it has just become incredibly cheap to generate content that is being published to the web. So, you can get any of the large language models like ChatGPT or Claude or any of these ones to just generate like millions of words in a couple minutes and it’ll cost you like pennies. And you can generatekeyword-stuffed articles on anything you want under the sun and publish those to the web.

And I think it’s still an open question of like, what happens to Google search in this world? Because we don’t quite know how Google’s going to respond to this outpouring of generated content, which is already happening. We have plenty of evidence that people are already doing this. But it means that if you search for a topic on Google that otherwise would’ve led you to someone’s personal website with their personal opinion on it… an opinion that is grounded in like a very embodied reality — their experience of the world, who they’ve read, who they know — you are instead going to… all the top results will just be generated content. It’s just going to be, you know, rehashed stuff out of language models. And that doesn’t mean that the content isn’t true or it isn’t accurate, right?

We have trained these models to actually be quite accurate, but there isn’t a human behind it. So, you can’t have a relationship with whoever’s writing these words. And while it’s more likely to be accurate and true, it still isn’t grounded in reality. When it comes down to it, those words could be false and but we have no way to validate that. And you have no way to check it, because you can’t contact the person who wrote it because no one wrote it. It’s like it was just generated text.

So, I think I’m very worried about our ability to connect with one another and like form relationships when everything you read on the web no longer has a human behind it, and how we stay grounded in like empirical scientific reality if there’s just this explosion of generated stuff, which includes lots of hallucinations. But we don’t know which content’s been hallucinated and which one hasn’t.

Jorge: There’s a flip side to this, which is that for people who are writing and learning and sharing in the way that you are doing… well, I’m going to speak for myself, I have found the LLMs to be useful collaborators in that work. I’ve written about how I don’t let the robots write for me, but I’ll bounce ideas off the robots.

Maggie: Yeah. Yeah.

Jorge: So there’s that flip side to it. I was thinking as you were describing this about the effect on human connection and just tying it back to what we were just talking about how your site integrates web mentions. Like, I could imagine that if the web mentions on your posts are picking up bot accounts, that might trivialize the responses in a way that reduces the value of the… maybe it introduces noise into the bat signal, somehow?

Maggie: Yeah, exactly. Right. Like, a lot of the really valuable responses I get to my posts are often… I mean, sure. People commenting through web mentions or a lot of it happens on Twitter, right? I’ll publish things to Twitter, and then really the comment section is usually really quite rich with insights. And it’s invaluable stuff because people will say like, “oh, well you haven’t referenced this thinker who’s really important.“ Or, “you missed this point.“ And you go, oh yeah, I did! You know? And you kind of go back and revisit your ideas and update.

And that’s a really valuable feedback loop as an individual trying to improve how rigorous your thinking is. Trying to holistically understand a topic; trying to figure out design solutions to real world problems. You need that feedback loop of other people. And it’s unclear what’s going to happen if you’re publishing and a lot of the comments coming back are just like… Twitter? Like the language model run Twitter bots are just sort of being like, “oh, nice article. I really liked point X,“ or whatever. But all it’s doing is kind of scanning the content and like making superficial summaries of it.

I mean, I definitely agree with you that language models are really useful tools for critical thinking and research. Like the lab I work for, this is what we’re exploring. Language models as rubber ducks, where you’re just talking to them to like get your ideas out, or getting them to critique your writing or brainstorm with you. I’m definitely in favor of that. I do that all the time. Like, I have a really strong prompting workflow at this point.

I think I only have a problem with people using language models to generate the full text. Like the whole article, right? Where they’re just like, “hey!“ You know? “ChatGPT, make me a thousand articles on topic X, Y and Z.“ You know? “Use these keywords, make it in favor of this political opinion,“ and then just like mass publish those. That’s not a human in the loop of like a thoughtful reasoning process, right? That is just like automation at scale.

So I think… yeah. There’s like a great way to use language models to like make us all better researchers and critical thinkers. And there’s a really negative way to just like flood the web with noise.

Jorge: It feels like we’re moving to a situation with large language models where at least some people are delegating learning and thinking to these things, right? And I see the practices that you’re advocating, digital gardening in particular, as a way of kind of learning in public.

Maggie: Mm-hmm.

Jorge: I had an interview with Nicole van der Hoeven a few episodes back where she talked about this idea of using tools like … well, she’s using Obsidian, but she’s also sharing in YouTube and she’s doing it as a way of thinking in public. And I guess the question more broadly is: why should folks be doing this? Like, I assume just from looking at the stuff that you put out, that you spend a lot of time working on your website. Why should someone be investing their time cultivating a digital garden in a world where there are LLMs that can think for you, right?

Maggie: Right. It’s a good question. I’ll first say, it definitely depends on your goals. But if you have any job that is at all what we would call knowledge work, which I assume is most people listening to a podcast called “The Informed Life,“ to be honest, you need to have like a set of beliefs and opinions, that you are bringing to every job you do.

And not even necessarily a job, right? Anything where you’re like, okay, I care about this, I want to work on it. And it’s some like meaty programming or design problem that you really think is worth solving. That’s really going to take some critical thinking. You’re going to have to be like, “okay, what’s been done in the history of this? What questions should I be answering? What are my uncertainties? How sure am I of my beliefs?“

These are kind of like core pieces of critical thinking that we all have to do, right? If we’re really going to be problem solvers in the world. And writing is one of the best ways to figure all these out and organize them and be like, “okay, I have to actually identify what like the critical questions are to answer. And then I have to write those questions down! And then I have to try to answer those questions.“ Right?

This is a process, and I don’t know a better process to go through than to simply write those all down and then try to answer them by writing stuff down. Obviously, you can use things like drawing and audio notes and there’s… you know, there’s all sorts of workflow stuff you’ve been doing here, but you really want to be very explicitly capturing your current understanding of a situation and your current beliefs about it to like be an effective agent in the world.

And the reason to do this in public? You could certainly do this in a private note-taking database and just kind of like write all this stuff up yourself. But doing it in public really raises the stakes, right? I’m a big fan of learning in public because I could write myself a really terrible answer to one of these difficult questions in private, and it doesn’t really matter that it’s a terrible answer. Like, who’s going to see it?

But if I’m like, “oh, I’m going to try to answer this difficult question in public?“ It’s like, oh no! Okay, now I have to like actually do my research. If I don’t, someone’s going to critique me. I think there’s a little bit of healthy fear in there of like, you could kind of get flamed on Hacker News if you do this badly. It makes you be more rigorous. It really makes you double check yourself and be like, “well, I’m making this claim. Is that claim really true? Do I have evidence for that claim?“

I’ve just found that writing in public in a way that doesn’t feel too high pressure has made me a much more critical thinker. It’s made me really do my research. It’s really made me learn my history and then that’s paid off in spades when I actually do have to solve a real problem at work. It’s like, “oh! I’ve researched this. I know this stuff. I can come to this discussion quite confident in what I want to advocate for because I’ve done my homework.“

Jorge: That is great. And as with your work, like I said earlier, just hearing you talk about it, it’s inspiring. So thank you for sharing with us. Thank you for putting out the bat signal — and I hope that it keeps drawing more interesting people your way. Where can folks find out more about you?

Closing

Maggie: Sure. Well, maggieappleton.com is the said digital garden that we’ve been discussing on this podcast. I’m on Twitter a lot. My handle is @mappletons; my DMs are open. I love it when people message me. I love just having open Twitter conversations. I think that is probably the best way to reach me, I feel like.

Jorge: Fantastic. I will include all those in the show notes. Thank you Maggie, for sharing with us today.

Maggie: Thank you so much for having me on. This was a really wonderful chat.