Towards a Metaphysics of Worlds
Meta-commentary on the most fun talk I’ve done in years
In November last year, I did a talk at the Autonomous Worlds Assembly in Istanbul (part of DevConnect, a major Ethereum ecosystem event) that’s probably my personal favorite among talks I’ve done in recent years. It synthesizes several themes I’ve been exploring for nearly a decade, starting with a 9-year-old post, On the Design of Escaped Realities. I further developed the line of thinking in a collaborative blog series with artist Ian Cheng called Worlding Raga in 2019, and a short story in 2022, The Map. Most recently, the writing I’ve been doing in this newsletter in the Protocol Narratives series fits right into this line of thinking. The tldr of this talk is: how to investigate the nature of worlds using a mix of traditional philosophy concepts and the computing concept of an abstraction leak as a primary investigation tool.
The AWA organizers just put the video out, so this week’s newsletter is the video plus some notes and highlight points (I don’t have a transcript, unfortunately). Here is a link to the slides. This may seem like a glimpse into a very obscure and nerdy subculture for many (most?) of you, but I think something very important and interesting is brewing in this scene and more people should know about it.
First, the context. “Autonomous Worlds” is a specific idea inspired by the Dark Forest onchain game, which in turn is inspired by the Cixin Liu trilogy. The game play involves discovering the existence of inhabited planets on a simple 2d grid universe, and attacking them. The game pioneered a whole mini-generation of technologies in the Ethereum ecosystem. “Autonomous worlds” is the idea of extending that technology to host entire virtual cultures and communities onchain. Perhaps even virtual nations and such. There’s definitely a “what are you smoking” aspect to the idea.
The creators of Dark Forest (the game), two talented young tech visionaries I’ve gotten to know over the last year or so, Brian Gu and Justin Gilbert (the Justin I refer to a couple of times in the talk), went on to found an unusual and somewhat stealthy decentralized research organization called 0xParc (a play on Xerox Parc and the convention of indicating hexadecimal numbers with a leading 0x, which is a popular branding strategy in the crypto world). They are driving a lot of fundamental research and engineering on zero-knowledge proof (ZK) technologies, as well as this more applied research into autonomous worlds, which partly builds on the ZK foundation, but also on a bunch of other convergent techno-cultural trends (such as the retreat to cozywebs for example, from what Yancey Strickler called the Dark Forest condition of the public internet).
ZK technology is some of the most genuinely exciting stuff I’ve seen in my career in technology. If I were thirty years younger, I’d be diving right in (I’m frankly too old and not smart enough to do so now). The roots go back to the 80s/90s, but what’s happening now is turning the abstract ideas into a technology stack you can build with. It’s effectively a new sort of digital building material. The 0xParc people call this whole emerging domain “programmable cryptography.” ZK tech goes together very well with blockchain technologies (it’s like peanut butter and jelly), but is an independent vector of evolving technology. The blockchain-specific relevance has to do with things like zk-rollups (a way to do Layer 2 chains on top of Ethereum with various benefits), but ZK tech is very relevant to all sorts of contexts.
If you’re an engineer looking for a new frontier, and can stomach some challenging cryptography math and very weird programming techniques, you should look into ZK. I had the opportunity to sit in on a 0xParc workshop on it last summer, and though I like to think I am good at grokking almost any technology to first order fairly quickly, ZK tech was (and remains) extremely tough for me to wrap my mind around. But if you can master it, the payoff is worth it. From my limited exposure, it looked a little like Verilog, the standard Hardware Description Language (HDL) used in semiconductor design. Though it is software engineering, it smells like circuit design. If you have the chops for it, you can explore the open-source tools available already. It is a weird sort of very counter-intuitive black magic. The intersection of ZKP and machine learning (“zkml”) is particularly fascinating and some truly magical tech might emerge from that collision.
My talk was part of a larger context-setting session at the start of three days of pretty intense idea flows. I highly recommend watching some of the other talks, which are listed in this thread. In particular the talk by Eve Online founder Hilmar Veigar Pétursson, who spoke just before me (hard act to follow) might provide a natural entry point from traditional gaming and world-building. Besides the talk, I was also on a panel on decentralized narratives. But the most interesting part of the conference was probably the demos of all the weird second-generation onchain games that people are building. I particularly liked the games being built by a group called Moving Castles. I’m not really a gamer, but even I can tell that something very radical is going on here, relative to the history of games going back to the earliest arcade and console games. On a conceptual level, this conference probably featured the most sophisticated thinking and ideas around world building I’ve ever come across.
A sidebar of sorts. There is a moveable-feast burning-manesque thing in the Ethereum ecosystem called Zuzalu, which operates as a sort of pop-up city in various places, serving as a cultural container/incubator for many of these technology things. A rudimentary technology demonstrator product powered by ZK tech called Zupass (a kind of ZK-based ticketing system crossed with a trading card game and a way to do NFT-like things without blockchains) was born out of stuff 0xParc people did in collaboration with Zuzalu. There’s not much to see at the zupass link, but if you meet someone technical who went to DevConnect last year, ask them to show you Zupass, in particular the Frogcrypto trading card game. It gives you a hint of what sorts of things are possible with the technology. Initially I dismissed it as an annoying gimmicky thing, but once I got what was going on under the hood, all sorts of fascinating possibilities and use-cases began to suggest themselves.
I’m still making sense of all this myself, so I’m probably not an ideal guide to the scene, beyond just pointing to it and describing it vaguely.
I suspect I ended up getting drawn into this purely by accident, by way of some of my writing having some sort of resonance with this developing idea space. At the event, I felt a bit like a random old guy who had wandered in off the street into a scene full of much younger people doing bizarre things that somehow still felt vaguely familiar and rhyming with stuff I get into.
But I suspect, as I think about this more and develop better intuitions about it, I’m going to find some really interesting bunny trails to go down.