I made this new section on my substack because I wanted a place for my lab notes while working on my research under epistemic garden.
The gist of what I’m doing with epistemic garden is exploring the value of applying AI to social data. I’ve been researching and building prototypes.
I’ve been limiting myself to community archive data because it comes from my own circles which I think are exceptional and produce valuable discussion, plus I’m familiar and intrinsically interested in them. The twitter schema is also pretty stable and simple, which makes working with it smoother than e.g. trying to integrate data from multiple sources.
I built the community archive last year and then I asked for money to build tools to map how ideas spread online. I was lucky enough to be funded and so created epistemic garden as an umbrella org over the community archive. We also published a browser extension that keeps the CA up to date with live data (monitor it here).
I was originally going for “open memetics” as an org name, which is catchier, but I’m interested in a broader scope. Not just how ideas spread, but also how to use AI on social data to help people make decisions and act together. This can take the form of mapping discourse, matching people, or - yes - doing research on idea dynamics.
I keep meeting people working on this cluster of ideas. They seem to “get” one another, not like they’re uncorrelated verticals. I wanted to get all the best ones in a room so I hosted the unconference on AI for collective sensemaking in New York. 1
It was very successful. It sparked new collaborations between peers who didn’t know about one another before, and I got a lot of feedback about how cracked and at the leading edge all the speakers were.
The conference going well made me feel encouraged to keep building the field, writing about it, making connections between people. I may or may not be planning a week-long summit in Portugal next year…
While I was in NYC I was fortunate enough that two students from the Fractal Tech bootcamp were interested in doing projects for the community archive. Noa did some experiments building real-time tweet embedding infrastructure and published a social graph exploration website. Boris reimplemented the Birdseye frontend in React but it’s not live yet because it still needs to be hooked up to its backend on Modal.
Meanwhile, IaimforGOAT (community archive homie dev #1) published a real-time embeddings pipeline, with an associated frontend for semantic search!!
My former colleague Alexandre and I have been trying out this collaboration where meet every day, take notes on a tldraw board, and pair the whole morning on research or prototypes. It’s been rewarding and I’ll write more about it in the next updates but for now I’ll say we published a website called Bangers that lets you see the best tweets of the community archive canon and explore their threads, quote-tweet chains, and semantically adjacent posts in an Andy Matuschak-inspired interface.
This was a recap of what' has been happening with epistemic garden and the community archive, thanks for reading!
I don’t love the word “sensemaking” but I called it that because I couldn’t think of better names that fit the broader cluster.