Conversational AI: A Tool for Knowledge Development

What you will learn

• Utilizing technology for efficient note-taking (02:28)
• Personal knowledge management with Roam, Notion, Logseq, and Obsidian (03:47)
• Using Readwise and Todoist for workflow and note-taking (04:59)
• Exploring connected note-taking apps – comparing Obsidian and Notion (10:09)
• The power of externalizing thoughts through knowledge management tools and AI (12:46)
• The future of personal knowledge management (13:32)
• Externalizing knowledge to make it accessible and effective (15:20)
• The potential of digital assistants and notetakers to assist in managing information and improving productivity (15:45)
• How generative AI helps users dive into more beneficial information (17:24)  
• Features of the “Ideal Tool” that provides automatic AI summaries for sources (18:38)
• The need for the ability to discern authorship and authenticity (22:47)
• The probable emergence of personal tutors and new forms of libraries (24:13)
• Tool recommendations to thrive on information overload (27:20)

Episode resources

• Roam Research
• Notion
• Logseq
• Obsidian
• Readwise
• Todoist

Transcript

Sam Barton: Hey, great to be here, Ross.

Ross: You are passionate about trying to work well with lots of information and manage your knowledge and so on. How come? Where did all this start for you?

Sam: I’ve always loved technology. I’ve always tried to keep abreast of all of the latest updates and just check out new tools and how they work. I think I stumbled into it in a way because I was using all of these different tools, and they had all of these capabilities, and as a consequence of using these tools, I started amassing a collection of notes rather easily because of the automation that some of these tools made available. Suddenly, I had lots of data to work with and I thought, this data can be put to use in a way.
I was working on various different things. I always work on various different things, for better or for worse, so having a way of going through the information that I have and organizing it in such ways that it’s actually useful. I was running a podcast, I was studying, I was working, so I had lots of balls in the air and I needed to find a way to keep those balls in the air.

Ross: What was the first software that you found useful on that journey?

Sam: That’s a good question. Roam Research captured my imagination. It got me very excited for this world of personal knowledge management, though it wasn’t as useful as I was expecting it to be. Notion is an app that I have been using consistently for the past few years. Even though it hasn’t gripped me in the same way, as some of these other tools, I have found myself using it the most for getting work done and tracking things. In the past, and currently, I use tools like Roam, or if not Roam, now Logseq or Obsidian. We can get into those later if you want. I’ve been using those for managing notes like book notes, podcast questions, and a whole array of things. Basically, stuff that I capture online, and then Notion is where the production happens.

Ross: How about in terms of the idea captures? You always come across something cool or interesting, which is stimulating, but you don’t necessarily know where or how, then where would you put it?

Sam: My workflow at the moment is I take advantage of Readwise. Readwise is an app that does a variety of different things. It’s evolved recently. The big thing is it allows you to capture content in the various areas that you come across it online, on Kindle, and elsewhere. It deposits that information automatically into a note-taking app of your choice or in the Readwise app itself. I had a Twitter addiction, I deleted it recently. That was a good and bad move, I think, because I’m less plugged in. But what I would do is if I came across anything interesting, I would just send it to Readwise, and I would have it tagged with #inbox so that I would know when I go into my notetaking app, I need to process this, I need to put this in an area that makes sense. I’ve also used Todoist to play that role as well. If I come across something, I would just throw it quickly into Todoist and tag it with the relative tags: work, podcast, or whatever, and I’d get to it later. I will say that my process is not perfect and it definitely needs a bit of work.

Ross: Nobody’s process is perfect.

Sam: Yes, I still have lots of frustrations with mine. For one, remembering to process your inbox is a chore. Maybe it’s one of those things that you just have to do.

Ross: Where’s the inbox?

Sam: My inbox was Logseq. Logseq is an open-source free version of Roam Research. That was where I was initially. I’ve moved to Obsidian recently because with these connected notetaking apps, you can hashtag or link items, do a quick search, and filter for many different things, and that’s one way of organizing and getting access to content. But folders are still very useful I’ve come to discover. Being able to just put something in a folder and access it in some hierarchical manner is very useful. From what I’ve learned, or from the research that I’ve done, Obsidian is the best app that allows you to make the most of both worlds.
You can organize things hierarchically in folders, but you can also, with tags, surface things when you need them. I’ve started moving to Obsidian. I do have some gripes with it but it’s high customizability and getting into the weeds and owning your data. I contrast it with Notion, where everything’s set up in a very structured way, but you don’t own your data so it’s not as future-proof. I’m currently in that space of tossing up whether or not I stick with Obsidian and just let that be my home for everything or I just bite the bullet and go with Notion because the returns initially will be greater. But in the future, having complete control over my data is the way to go.

Ross: I use Obsidian, Notion, and Google Docs amongst other things. That’s always one of the issues, either you’re choosing a central platform, a single platform for everything or you’re making a decision about what goes where. At the moment, there’s a logic as in Obsidian is for me more of the idea. Here’s the idea, this little snippet, a reference or an image, or something which I can go, organize and get access to whereas Notion is where the structure happens. As you talked about Notion being that place where work gets done. I don’t think Obsidian is a work-getting-done tool. Though I’m sure some people put it to a good effect in that way. Just working out the boundaries as to if no one told us at all, then how do you continuously rebalance what goes where?

Sam: It frustrates me because having everything in one place is very powerful, just being able to manage, let’s say, all of your resources, reference notes, and everything, and then just being able to seamlessly integrate that. Let’s say I’m writing an essay, a workflow for me would be to read various books or PDFs and highlight them. My workflow allows for all of those highlights to be automatically added to my second brain, so Obsidian or Logseq. A really useful workflow is just being able to take all those highlights and then just put them into my working document, so that I can write with all of the content right there. But unfortunately, it doesn’t work like that. I end up having to get all those notes there, and then move to Notion, and then port everything over there and try reworking it.
The reason why that’s a problem. It doesn’t sound like a problem. That’s normal and that’s how it normally happens. But the benefit of these connected notetaking apps is that you can see the providence. Where did that note or insight come from? If I’ve got a highlight, this highlight came from this book in this section, and it’s really easy to reference and look at later. In the future, I could say, where are all the places that I have referenced this highlight in all my work? And maybe there are 10 places. Now that’s very useful because you have the graph of knowledge that expands across time and becomes more useful.
There’s this positive feedback loop between data. But when you just move it over to the workspace, you lose all of that. I would really like Obsidian to play that role for me. I see it as a niche hacker-ish tool. It’s a tool for hackers, and people who like to tweak things, and it’s not user-friendly enough for the sorts of things that I would like to do with it, given the amount of time that I have, but I do think that it’s in the right direction.

Ross: I recently gave a keynote to 250 innovation professionals and I asked at one point, ‘Who’s heard of Rome Research or Obsidian?’, and one person put their hand up. She’s an avid user.
It’s interesting that these are, as I’ve referred to them before, geek tools, and that’s the challenge, is that they are great for geeks, but we need tools that are more accessible to others that don’t necessarily like going into the weeds like that. But pulling back, we’ve used the phrase ‘personal knowledge management.’ For me, these note-taking tools, I think of them as knowledge development tools. As in, how do they help us be more knowledgeable, to build our mental models, to understand better? I’d love to just pull back from the specific tools around thinking how is it that you, as your knowledge, your insight, your perspectives, what is the big frame for you in being able to support that, being more knowledgeable, essentially?”

Sam: Yeah, it’s a great question. I don’t have a very good answer for it because my brain still acts as the main brain.

Ross: Yes, as it should be.

Sam: Though I do have a strong desire to externalize a lot of it so that it’s accessible not only to me; but, especially today with the rise of generative AI tools, the more I can get out of my head and digitized, the more effective my thoughts will be, if that makes sense. Because I’ve taken things that are in here, like things that I’ve found myself, and put them into a knowledge management tool, and then I can work with the AI to explore all of these ideas and refine them and do all that stuff. The frame that I have around this at the moment is we’re just at the beginning.
We’re still at step one. Maybe like pen and paper was 0.1, Microsoft Word was 0.2. We’re still at the very beginnings of personal knowledge management and working with external technologies to really amplify. Amplify might not be the right word but to immerse ourselves in this information and work with it in a more collaboratively way so it’s not just my brain throwing things, and arguments happening in my brain, but I’m having a conversation with the knowledge that I have come across, but also have synthesized and put together in my own words. I don’t know if that’s an answer to your question, but that’s how I am thinking about it at the moment.

Ross: No, I think that’s a really important point as the times when I’ve gained the most insight is being in conversations. Sometimes things I said, not what the other person said, were what we both said. I say things like, ‘Oh, wow, I’m interested in what I just said,’ or then indeed you’re having a conversation with someone else, and they’re saying something else, and you get a refinement or addition. When you’re sitting and reading and making notes and pulling things together, yes, that’s valuable. But in a way, most of the really big insights happen in a conversational frame. As you say, now that we have AI that can potentially play that role well, that could be an extraordinary tool for knowledge development.

Sam: Yes, just on that point, something that I’ve been thinking about; I’m very talkative, and I love talking to people. One of the reasons I started the podcast was like, “Oh, well, maybe I can make a living just chatting with people.” That sounds spectacular, chatting about ideas. I’m not very good at sitting down and writing notes and doing all sorts of things but I love chatting about things like this. With the rise of generative AI and transcriptions and all that, we could actually have conversations like this, the transcript gets automatically dumped wherever, and maybe in a year or so, the notes that we would have taken manually will just get pulled out by AI and tagged and all sorts of wonderful things would happen to them so that I don’t need to do anything.
It reminds me of back in the day, in ancient Greek philosophy days, 2,000 years ago, none of the philosophers, the big dogs, weren’t writing things down. They would have students who would recall things and take notes and everything. They said oh, no, writings, we’re not about that, we’re just going to be talking about it. They didn’t have the patience to write things down. I’m not comparing myself to them, maybe just in terms of my laziness. But now we will have these digital students, or digital notetakers, and digital assistants, who’ll really help us with all that now. I’m here for it.

Ross: Yes. How else have you found the generative AI tools to be useful in your knowledge? The tools are very pragmatic in terms of being able to use them for all kinds of work, but how have you found them useful in terms of thinking better?

Sam: It’s a good question. I would say that they haven’t been as useful yet, though there’s a lot of promise. There are some ways in which I’ve used generative AI to understand things like just copy some text and get the AI to explain it to me in a different way, reframe it, or maybe combine ideas in a certain way. That’s been the most useful way so far. Knowledge is tough, but inputs are easy; if that makes sense. Inputs precede knowledge. But what the AI will do, the benefits that we’ll see is these generative AI will allow us to triage increasing amounts of information so that we can identify what we should dive into in more detail and extract value from. So right now, if I come across something that I want to read later, I would tag it, throw it into my inbox, and maybe read it later. To read it later, I actually have to read it. But now I can click summarize and maybe get a rough summary of it, and have an idea.
A tool that I would love to exist right now is something where I could put in all the sources that I would like to keep up to date with, and every time they create something new, let’s say, it’s a blog, every time there’s a new blog or a new podcast, or if it’s a Twitter personality, their Twitter threads, or whatever, if I had an automatic AI summary generated, and I had a feed of summaries that I could then say, ‘Oh, actually, that’s very interesting. I’m going to go and dive further into that,’. Basically, increasing the, I wouldn’t say signal to noise, but just extracting the key insights from each piece of information so that I can process more and dive deeper into things that are worth going, diving more deeply into, because there’s so much information being created today, and it’s a bit overwhelming.

Ross: Essentially saying, you work out what the sources that you think are promising, and then you get the system to provide summaries or just key points and things you’re interested in from each of those so that you can go and follow each of those if you think it’s worth delving deeper into.

Sam: Yes, exactly. One tool that’s doing this better than anything else at the moment is Readwise. Readwise does have a reader app. You can send PDFs, online articles, anything you come across online, Twitter threads, YouTube videos, and podcasts now, and you can read it, view it, and consumes it in the Readwise app. They’ve got something called the Ghostreader. You can just click a button and it generates a summary. That’s a step in the right direction. I just want an interface on top of that, with the summaries already generated so that I can filter information more effectively and dive deeper into things I think are worth exploring.

Ross: Yes, I see that tool to be extremely valuable. But I also wonder it still just provides the text and text summaries. There’s still this brain, the wetware of our minds. Obviously, there are some images and other types of content, but a lot of this is text, and maybe summarize text, we can delve deeper, but there is still this step between how it is the different pieces and elements that we come across, and be able to build a lattice of knowledge from that. It’s a step beyond, again, where we can find the best things, but also then find how it is we pull these together into our lattices of knowledge.

Sam: That just harkens back to the discussion of Obsidian versus Notion, owning your data versus not owning your data because if you have all these summaries, and you have all this information that’s tailored to your desires, to your interests, to your needs, then you can see them all connected together and interact with them in this spatial way. I’m just thinking of the graphs that Obsidian puts out or the Obsidian, Roam, and Logseq, so you can just see all of your different notes connected and how they’re connected. I think that’s a very useful development. But right now, there’s just too much friction involved with that, but probably not for too long; maybe by next week, seeing the pace of this generative AI explosion, we’ll probably be there in not too long.

Ross: Just looking at that broader space, the intersection of personal knowledge management or our ability to make sense of information and build our knowledge with AI, is there anything else, in particular, you’ve found exciting, which you have seen already or where you see the potential of that coming? Where’s the edge of where you see possibilities now and coming up soon?

Sam: I’ve got two answers to this question. The first thing that springs to mind is actually a negative. We now have the capacity to generate very convincing content at scale, and we have no way of discerning facts from fiction. COVID has made this very apparent. Now it’s already happening, but over the next five years or so, we’re going to see deep fakes, generated audio and video, have quite a big impact on people’s worldviews, and collective truth will just become a myth, at least in the near term. What I’m excited in this space is the development of decentralized identity tools, a weld coin is a good example.
They’ve been in development for a while but they have led to building a decentralized, privacy-preserving identity solution so that people can prove their humanity. What we need ASAP is the ability to discern; if I’m reading someone on Twitter, if I’m seeing a post from someone on Twitter, I know that they’re a person and I know through some cryptographic key that they’ve signed authorship through some cryptographic signature, they were indeed the origins of that post. That’s something that I’m excited about. Well, excited is not the right word, I see that that’s something at the front of my mind.

Ross: It is what you’re excited about?

Sam: Yes. It’s communities of people who have shared interests coming together to create libraries on certain topics. An example that springs to mind is the whole John Vervaeke space, the Meaning crisis, philosophy, and civilization. There’s a lot of conversation happening in a distributed way on Twitter and YouTube about our collective stage, the challenges that we face, and what sorts of changes, in the way that we operate individually and collectively, maybe we need to undertake to be able to deal with the challenges that we face. There’s no Wikipedia page for it. There’s no ongoing collection of information that I can go into and just review that’s been updated in real time so that I can keep abreast of the developments there.
These tools, with some curators, we’ll see new forms of libraries appear with people who are paid librarians, who are AI wizards, who will update these knowledge resources, these libraries in real-time using artificial intelligence. There’ll be digital gardens that tend to themselves in a way and emerge. That’s something very wonderful and exciting. The other thing is personal tutors. That’s the other thing, just being able to have a tutor on anything and it being tailored to myself. I can’t wait for that.

Ross: Just on the point around the curated gardens of knowledge, I think the last episode with Ida Josefina, with their product Sane, which is not yet publicly launched, but looks very much to the intent of what you are describing and it looks very promising and interesting. That human curation element is very interesting, sane.io, I believe.
Pulling back, you have been very effective. One thing about you, Sam, is you keep on trying lots of tools all the time and learning what is best, which is not just what the software is, but how you use that. There are plenty of other people that don’t either have the time or spend the time to go and explore and try things and see how they work. I’d love to pull back to just get your recommendations for anybody that’s immersed in a world of information and looking to build their knowledge effectively. Not just talking about software, but just more generally, what are your recommendations? What are the things you would advise to someone to thrive in a world of overload, to amplify their cognition?

Sam: Yes, the first one is Readwise, particularly if you read eBooks, because you can just easily capture highlights seamlessly, and they just get saved. My recommendations are largely based on minimum effort.

Ross: That’s a good principle.

Sam: I would say, a recording app. You can choose whichever one you want. Otter.ai is one example or the Google Recorder, basically, an app that allows you to, whenever you have a thought, just click a button and capture the thought in real-time. Just speak and have a transcription generated so that you can leverage that later. The reason why that’s useful, particularly today, is if you have an idea, and you get it out, and it’s machine-readable, so you get the transcript; with generative AI, you can turn that into whatever. You could turn it into a blog post, meeting notes, or an outline for a book, you’re just limited by your capacity to work with these tools. It’s all about capturing useful information, so Readwise, a recording app; I would say a tool that integrates.
Let’s say capturing is finished, the last set of tools I recommend relating to curation. I have used Notion extensively and it’s continuing to get better and better, particularly with the integration of AI. I would recommend using a tool that allows you to integrate generative AI into your workflows seamlessly. Because we now live in an age where you can be 10 times more effective than your peers, if you’re a knowledge worker, this isn’t just being a programmer, it’s not about just being a 10x programmer now. If you know your way around tools, around this generative AI space, you can literally be 10 times more productive than your peers or 100 times if you’re incredibly strategic and creative. I would say find a tool, I think Notion is one of the best examples of that at the moment that you can just use to do what you do normally but throw some AI in there as well. Coda is another one to I would say.
I would love to know Coda in more detail. But I’ve spent enough time with Notion that I might just wait for them to catch up. But if you’re just starting out, I would look into both. Consider what your needs are, and then make a decision. Just look at a YouTube video Coda versus Notion and make the decision. Those would be my top things.

Ross: Just beyond technology, any attitudes, habits, behaviors, or anything else which you think is important?

Sam: Yes, openness. So, if you can cultivate an open mind and a propensity to try new things. We are in the most atypical time in human history, possibly. We are going through a massive change literally as the days go by things are changing and it’s not just generative AI, we’re all very excited by that, but we have this confluence of technologies coming together that are all accelerating at a relatively exponential rate and they all feedback on each other. This is a positive feedback loop. That’s why I call it the Law of Accelerating Returns, returns to something. Things are going to be changing very quickly, they are, and they are going to continue to change. To be able to engage with this turbulent world effectively and not going to not get overwhelmed, we need to try to not be rigid. We need to try to be as flexible as possible but within reason. That’s my advice, just be open to new things.

Ross: Open but not too open.

Sam: Yes.

Ross: One of the problems with openness is that you can be open to all sorts of incorrect information, for example, but that’s in Thriving on Overload. That was one of the key points I made that openness is not just extraordinarily valuable. It is something that we can cultivate in ourselves, and it pays off in spades in a fast-moving world.

Sam: Yes, definitely.

Ross: Anywhere where people can go to find out more about what you do?

Sam: Yes. I only really use Twitter these days. I’m @Samhbarton on Twitter, and my website is the same thing: samhbarton.com.

Ross: Awesome. Thanks so much for your time and your insights, Sam.

Sam: Alright, thanks for having me, Ross.