Twnstr Manifesto
2026-11-11 · Alex Maestretti, CEO
The diffusion of artificial intelligence appears to be the most impactful change we will see in our lifetimes. Assuming we avoid being turned into paperclips the coming societal impacts are large and difficult to predict. If the past is prolog: 1) technological progress is effectively irresistible in a global society and 2) while technology can empower individuals, on balance it has consolidated power in institutions. The question that animates us is where can we create individual and small group autonomy within a framework of institutions that promote common humanity?
We see an opportunity for artificial intelligence to be a convivial tool that empowers and uplifts, but also a present day risk that jobs are destroyed faster than abundance is realized.
“Idealists will speak of the reduction of the work week. But this reduction can only be effected when equivalent technical improvements are produced in all fields.”
— Jacques Ellul, Technological Society (1954)
The power to create software has been democratized giving small businesses and individuals more autonomy, but also rendering some jobs obsolete. If efficiency gains are uneven - eliminating backoffice work but still bottlenecked on physical steps in the value chain - there will be a near term regression of quality of life for most. Autonomic manufacturing that allows an individual to create their physical reality moves us through that risky moment.
In many ways this is a return to earlier times where people built their own homes and forged their own tools, but this time maintaining the quality and safety standards we encoded in central institutions. It is difficult to build your own home under building, electrical, and plumbing codes and credentialing. It is difficult to build your own tools when the details of metallurgy and machining take decades to master. However, if you have an intelligence tool that knows those things, and responds to your intent, then as we have seen with software: custom solutions to your specific needs out perform general solutions required at scale.
We need to leapfrog existing manufacturing approaches that optimize through scaling and consistency, and develop autonomic manufacturing that responds to human intent with finished goods. The path to this runs through stronger general models with better reasoning about the physical world and manufacturing processes (which the frontier labs are creating), likely enhanced by specialized models and robotic policies within the tools to sense and react to material and environmental variation. These models will control physical embodiments across form factors enabling vertical integration from raw material extraction through finished product delivery.
The future is a broad collection of small scale manufacturing shops doing high mix low volume work with fast iteration and delivery similar to how software is being built in the age of AI. Shops will be staffed by tight knit teams applying human creativity, coordinated across an agent economy to manifest raw materials and electrons into finished human intent. We were meant to be creators, not consumers. This is a return to our purpose.