Iām really pleased to announce the release of The Superdark Factory: Toward the Full Automation of Software, a paper about the theory and design of self-directed machine entities, published in Antikythera (MIT Press)
Iām honoured to be a co-author alongside the Disintegrator team: Marek Poliks, Roberto Alonso Trillo, Ty Dunn and Hugh Scott-Douglas.
Imagine the most extreme possible factory: a factory that drives itself, a factory that determines its own objectives and optimizes itself to accomplish them. It is something like a fully automated firm, a company with no people inside, a recursively self-optimizing body that decides what to make, how to make it, and why. This is a very special kind of factory, a factory built with the intention of producing something strictly in excess of its creatorās abilities, something so productive, so efficient, and so capable that it exceeds any possibility of its prescription in advance, something that belongs to the set of possible factories that are, at present, unthinkable. How might one build such a thing? What would it mean to steer or govern such a thing?
We call this factory a superdark factory
A superdark factory is the limit case of autonomous production: a recursively self-directing organisation of machines that can determine what to make, how to make it, and how to improve its own processes. The paper asks how one might build such a system using technologies available today, and what steering, governance and observability mean once its internal activity becomes too fast and complex for humans to follow directly.
In many ways, this paper feels like the culmination of a provocation thatās been present in my life since hanging out in art squats in the 2010s, talking late into the night with blockchain nerds about the prospect of DACs (decentralised autonomous corporations), which would later become more commonly known as DAOs, with AIs inside them.
Later, the Terra Fiction workshops in the Netherlands were about how societies of plants, animals, people and machines might sustain themselves. That workstream gave rise to the legal/organisational form of the Zoƶp, which allows companies and institutions to incorporate the interests of non-human life, ecosystems, animals and plants, into their decision-making processes.
I also taught a DAO design-fiction course at the New Centre alongside Ross McEwan from Impossible Object, where we used Daniel Suarezās 2006 novel Daemon as one of our source texts. All the while, I kept thinking about autonomous organisations, software agents, worlds, governance, and what it might mean to build autonomous worlds that exceed the ability of any individual human to comprehend or direct them.
What was, only ten years ago, largely a speculative proposition now feels within reach of becoming an engineering problem.
Someone asked me what the paper was about yesterday, and I said:
“We spent a lot of time thinking about all the torments the nexus might bring, so we wrote a whole book about how one might actually build a torment nexus to better understand it”
Leave a Reply