Welcome to this update from New World Same Humans, a newsletter on trends, technology, and society by David Mattin.
If you’re reading this and haven’t yet subscribed, join 26,000+ curious souls on a journey to build a better future 🚀🔮
I sent the first in this series of Postcards from the New World back in December. Now spring is arriving here in UK; time flies.
The subject of that postcard was LLMs and the rise of simulated human societies. It looked at the work of Joon Sung Park, a PhD student at Stanford University, and still my favourite AI researcher.
Park and his team created Smallville, a sandbox-town peopled by 20 LLM-fuelled personas. They hit ‘go’ on that simulated mini-society and watched as the inhabitants made friends, wrote letters to one another, and organised a birthday party.
This week I discovered Altera, an AI startup playing in a similar space. It is led by Robert Yang, a former assistant professor at the MIT Department of Brain and Cognitive Sciences.
A few months back Altera launched Project Sid (its name is a hat tip to Sid Meier, the creator of the iconic Civilisation video game). They poured 1,000 autonomous AI agents into a Minecraft world, and sat back to watch what unfolded.
Altera calls this world, ‘the first agent-based civilisation’. Yang and his colleagues watched as the agents learned to harness the natural resources around them, developed their own economy, and even evolved a culture and religion.
To begin with, the AI villagers had nothing. They collaborated to gather hundreds of items from the environment: tools, minerals, and more. They established a marketplace, and agreed to use gems as a common currency, to facilitate trade.
Altera have run numerous Project Sid Minecraft worlds since the first. In one simulation, the villagers became worried about some among their number who had gone missing. So they ventured out into the night, to light torches intended to guide the missing villagers home.
I find these simulations endlessly fascinating. But watching these AI villagers fashion their own proto-society, I realised something new.
Back in the first postcard, I said AI-fuelled worlds of this kind will enable us to simulate complex human collectives of all kinds.
Now I realise that they’ll have a dual power. Yes, they’ll help us to understand the nature of human societies. But they’ll also give us an insight into societies, and economies, comprised entirely of superintelligent AIs.
And we’re going to need that insight.
We’re on the eve of a monumental shift. It’s a shift away from an economy centred around we humans: our labour, production, and consumption. And towards one in which AIs produce many goods and services, and autonomously transact with one another.
AI-generated social media influencers will hire AI-fuelled marketing services. AI agents will provide legal services to startups that are run and staffed entirely by AIs. AI agents will launch meme coins, promote them, and then sell them to other AIs at a wild profit.
This is the end of the human-centred economy. And the birth of post-human economics. And we can see glimmers of it now, via initiatives such as Project Sid.
When we watch Altera’s 1,000 agents run wild inside Minecraft, we’re being allowed an early glimpse of an economy inhabited only by AIs. And we’re being allowed, too, a sense of how it will feel when we are no longer the primary actors. When we’re standing on the edges, looking in.
Now scale this up to 1 million AI agents; or 100 million. Imagine the complex and nuanced emergent behaviour we’ll be able to observe.
Which raises, of course, another question. When does this kind of virtual world stop being a simulation, and start being a society in its own right? I think we’ll see virtual worlds inhabited by millions of AIs, trading with fiat currencies and crypto, and building GDPs that exceed that of real-world nations.
Perhaps those AIs will even create simulations of their own, to help them understand us. In this way, the simulators will become the simulated.
I’ll keep watching. Until next week, be well,
David.
This was #9 in the series Postcards from the New World, from NWSH.
Okay, why would we want a post-human future?
Btw, do check out the great futures work by Parsons & Charlesworth in Chicago for their Venice Biennal work on the Catalog for the Post-Human
https://catalogfortheposthuman.com/
Parsons & Charlesworth are continually let down on how much their futures research has become real.