Andy Matuschak
patreon
Bringing ideas into your Orbit
🕑 Added 2020-05-27 18:07:47 +0000 UTCAs I’ve mentioned in recent posts here, I’ve spent these last few months building infrastructure which I hope will help me (and others!) explore a wider set of ideas around systems like the mnemonic medium. In some real sense, it feels like I’m building my research lab!
One strange thing about building a system like this is that I have only a hazy picture of what I’m actually building. I build the thing to see, so that I can build the thing, and so on. My conception of the core structure is constantly evolving, and what follows is all extremely speculative! But I thought you all might enjoy hearing about how I’m thinking about it right now.
Computer operating systems have come with a predictable set of personal information management tools for decades: an address book, a calendar, an e-mail client, some basic note-taking function, files and folders, etc. These are structured differently from siloed “apps,” which typically aim to subsume some workflow from start to finish. This basic OS software is more general-purpose, each both a tool and a service, connected throughout the OS via API-powered integrations. You add an event to your calendar from an email, autocomplete a contact’s name within a chat app, save and open files to the same folder from many apps, and so on. More recently, we’ve come to expect these functions to seamlessly sync everywhere and to communicate with various web services.
What if there were an “OS-level” spaced repetition system (SRS)? What if, rather than living “inside an app’s shoebox”, as in Anki and other existing tools, prompts were framed more like files in folders—readable and writable throughout the system and by other services across the web?
Web articles could surface interleaved prompts, written by the author as in the mnemonic medium or perhaps by readers as on Genius / Hypothesis. You’d fluidly import these prompts as you read, just as your browser forms a history as you read.
Your PDF and e-book reader’s annotations could naturally be surfaced in this centralized SRS, rather than remaining siloed in some inaccessible sidebar.
Just as modern operating systems may create tentative calendar events or contacts based on chat messages or emails, the system may create tentative SRS prompts based on links you’ve bookmarked or phrases you’ve searched for repeatedly.
When you jot notes in your daily meetings, you could tag key insights with a special tag to surface them to this system—perhaps a future word processor’s formatting bar would include buttons for bold, italic, underline… and “to be reviewed.” I’ve built this kind of “personal mnemonic medium” for myself, and I’ve been using it for the last four months. I haven’t spoken about it much yet because it’s very odd, and I don’t understand it. But it’s fascinating, and—I think—promising.
But all these ideas become much more interesting once you think of SRS as useful for much more than memorization.
With the final chapter of Quantum Country, we’ve experimented with using spaced repetition prompts to help readers apply what they’ve learned, in addition to remembering what they’ve learned. In our personal Anki practice, Michael Nielsen and I have been experimenting for several years with using these interactions to prompt synthesis and reflection. Over the last year, I’ve been experimenting with using these interactions to support incremental creative work and for my reading queue.
For example, you can raise your smart watch today and say: “remind me to write about my idea that SRS could be framed as an OS-level service.” That’s a one-time reminder. But with this OS-level SRS, you could raise your watch and say: “remember to reflect: what novel contexts might benefit from an OS-level SRS service?” That would create not a one-time “to-do” but a prompt for repeated reflection over time.
This all calls for a broader perspective on spaced repetition systems. The typical image is of flashcards, used to memorize things, marked as correct or incorrect. But a more general framing is: with an SRS, you can arrange to repeatedly engage with some task over time, and the timeline can evolve with your actions. Some instantiations of this…
- memory: repeating recall questions; intervals expand/contract exponentially with success/failure
- procedural fluency: repeating short exercises with many variants; intervals expand/contract exponentially with success/failure
- writing prompts: repeating questions, inklings, and stems accumulate written responses over months; mark prompts as “fruitful” or “not fruitful” to expand/contract intervals
- habit formation: repeating short reflection prompts, possibly changing according to sequence, applied to your daily experiences; intervals expand by default but can be slowed
- reading queue: a queue of PDFs, tabs, books constantly revolves; if you ignore an item for a while, it’ll depart for a week, then a month, etc; say “I’ve read enough of that for now” and it’ll come back in a few days; say “not now” to an item, and it’ll depart for a few weeks, then a few months, etc…
- email inboxes, task queues, personal communication, etc
Central to what I’m imagining here is a daily practice habit somewhat akin to meditation: you open this thing up and engage with whatever microtasks it presents. You’ll work on your memory, maybe do some self-authorship with reflection questions, do some quick physics problems, some quick writing, etc. Then ten minutes later, the train arrives, you board, and that’s it for the day. The next day’s different.
In summary, this system is about giving you a way to bring ideas into your orbit. When something seems interesting, you can tie a string to it and throw it up in a lazy arc. It’ll swing back around at some point, but you’re not terribly concerned with when. You’ll give its string more or less slack over time. Floating above your head, then, is an ever-shifting constellation of inklings, facts, questions, prompts, obsessions. Every day you stare up at the slice of sky above you and respond to what’s there.
So at least for now, I’m calling this system Orbit.
————————
Your thoughts and comments are very welcome! As always, I’d like to thank you all for your generous support of my work.
I’d also like to mark a milestone: as of a couple weeks ago, your collective support now covers half of my mortgage! A year ago I honestly expected Patreon to be something of a token support mechanism, but thanks to you all, this now makes a meaningful difference in my life. Of course, I’m still in the red here, so this is not on its own a sustainable source of support, but it does help push out the timeframe for seeking other funding sources—which in turn, helps me focus on the actual work with minimal distortions. Thank you for that.
Comments
Andy Matuschak
That's very kind; thank you!
Just like Amir, I've been also struggling with a dozen different apps including Readwise to manage my knowledgebase and spaced-repetition routine. Like many others probably, I ended up with the obvious idea of combining & improving all these different tools I use into a general concept that if implemented as software enables me to easily capture, distill, amend, highlight & make additional notes and have them automatically reinforced. After skimming through marketing-style self-help non-sense for few months around the subject, now that I stumbled upon this Patreon and I see a professional is actually doing proper research on the subject with a much broader perspective, I've got really excited. I wish I could be of bigger financial support here.
Andy Matuschak
Thanks for sharing, Amir. This kind of fluidity is definitely a big piece of the goal…
Catching up on these articles late. A system I use is to leave emails in certain folders that I know I’ll get to in some indeterminate amount of time because I know it is important to read or follow up on. Email notifications about your articles are one type of those. I love the name Orbit and your explanation for calling it that. To add one anecdote to the system your proposing - I tried Readwise recently and gave it not one, but two chances, and it ended up underwhelming me. No math rendering support which a lot of the books I read need. No deep linking back into the books so I can discover context for something I’ve highlighted. No live synching service. My data felt siloed and as such became less useful to me. I want my notes to be fluid. I’ve been using Roam and it has been working well enough but I’m liking the evolution of the OS from your POV.
(Hi James) (lovely to see you here)
James Cham
(Hey Shripriya!)
I like that its simple and would flow well in notes. I often revise the text of my cards, however. I’d be interested to see if it’s possible to add the “heuristics or extra identifying markup” without cluttering up the notes.
Andy Matuschak
I describe my current syntax here: https://notes.andymatuschak.org/z4mAF1uBV96r72e4NjLcDaujEyTPGiUQJEj8C
Great! I’m always interested in the nuts and bolts. What types of syntax did you have in mind for the prompts?
I resonate a lot with Szymon's skepticism abour SRS! I'm surprised the wider context & aesthetic for the mnemonic medium in this post (specially the metaphor of orbits) has made me a believer! Like Szymon I keep multi-stage inboxes of thoughts which I then have to remember to keep reviewing. It's the challenge of my life to stay centered, accurately looking at all my inboxes, looking back & ahead while being present. I call it The Quest for Long Centers and have sometimes compared it to the layering of beats in electronic music. Now I can think of it as being in orbit, of having all my thoughts in proper orbit around me. :)
Evan Rocha
Great to hear :) I'm still pretty shocked by how much more I've retained from the Quantum Country essays vs. other things I've taken deep dives into without any SRS
Andy Matuschak
I think one of the most important design aesthetics of SRS—when it's working really well—is a sense of effortlessness. Not that it's very easy to answer the questions, but that no will is involved. You don't make decisions about what to do, you don't muster the gumption to do a particular thing… you just _show up_.
Andy Matuschak
Hi, Szymon! Thanks for writing and for your support. Indeed, my interest in using SRS-like systems for writing and reflection prompts follows past experiments similar to your own. I'd abuse GTD-like systems' deferral and recurring task features, or try to hang that stuff off weekly/monthly reviews. I couldn't make it stick.
Andy Matuschak
The mnemonic medium is the center of most of my research questions for the foreseeable future. The Anki-replacement angle is mostly important insofar as it enables experimentation with mnemonic media. I don't think an isolated Anki-like tool can become successful and widespread without the cultural knowledge of prompt-writing becoming substantially more widespread. The largest barrier to SRS adoption seems to be that prompt-writing is challenging and taxing; the mnemonic medium seems to be a great solution.
Evan Rocha
This sounds like what I have wanted Anki and other systems to be. Mainly reducing as much friction as possible to create the flash cards to make create more flash cards myself so I can get behind this (Orbit is a cool name too). I do think the mnemonic medium where content creators are creating content with spaced repetition in mind is end goal to me and this feels like the right first step to build the tooling to get there, but I hope that the mnemonic medium should stay the end goal
Hey Andy, I have to say I've been quite bit sceptical about SRS systems, mainly because I believe (believed? - not sure yet - occupying this middle state right now) that flashcard-like system for memorising things wasn't helpful for what I'm doing. It seems curious that I ended up with similar techniques, although not SRS-driven, approaching this from different perspective. I keep a "pondering prompts" document, which I look at in my morning pages writing sessions; I have a multi-stage system of inboxes for thoughts - quick capture, then "hatchery", then long-term storage. And then reminders to review, and re-review. I'm doing this with a basic markdown extension and custom cli tooling. I just keep on feeling that I have to somehow "keep the ball spinning" manually, when the Orbit system seems like a much more relaxed way to go about it, though it requires some "letting go" - I'd be curious if you had similar feelings, or did it come more naturally to you, since you've been using SRS for long time?
You rock dude!
This sounds completely awesome. Excited to try this!
Andy Matuschak
Thank you for this! This is a good sign that I'll need to choose a different metaphor, since I'm in violent agreement. :) Lots of businesses talk about building "an OS for X". Their aims are usually somewhat imperialist; they want to create an environment where you spend your whole day; their goal is for that platform to subsume other apps. I'm talking about doing the opposite: building "an OS-level system," unobtrusive and modest, integrated into your existing OS and its applications in the same way that your OS's contact manager and calendar are. More concretely, I've built something much like you describe. Orbit includes a Markdown extension which specifies how you can write prompts in any textual context. Then you can point the Orbit apps at directories (and, eventually, cloud storage) which contain relevant files; it scans them, tracks prompts it finds, and presents a review experience. It has its own internal synced storage you can use if you'd like, or not. In this sense, it works much like Ulysses and other Markdown editors. On the web, prompts can be embedded in pages via a standards-compliant Web Component. These are extracted into users' collections when they're answered. At the core of this all is an API which others can use to create alternative ways to read and write prompts. Many of the core primitives are actually designed to support a decentralized future—i.e. everything content-addressable, with no centrally-controlled server. Once the primitives solidify, I'd like to open-source and open-standard this stuff.
That's similar to the way I see it too. I think it should start as a markdown editor with some simple SR functionality. The reading list could be a simple list of links.
I really like what Orbit could do. In theory, I love the term OS, but in reality, it makes me a bit nervous. If we stay in reality for a moment, Orbit could be a set of apps or a set of plugins into apps. The vision of thoughts hovering over my head, ready to be pulled down or thrown back up onto a lazy arc is very enticing. Could Orbit be a markdown syntax which works in all existing apps and is a layer above them? That in itself would change how information moves between apps and back to us. The concept of a few minutes a day dedicated to thought exercises on the most promising streams is a world I'd love to see.
Andy Matuschak
Thank you, Ludwig! I'm currently targeting having a prototype client usable by Patreons in about a month, at that point functioning mostly as a simple Anki replacement. I'd like to have embedded prompts publishable in new texts by the end of July, but that timeline will depend on how generative some upcoming design explorations turn out. Will keep sharing! 🙇♂️
Andy, you’re taking this into a very interesting direction. I really enjoy reading your explorations and am looking forward to when you’ll have more to share! (Maybe even prototypes? \o/)
