One problem.
We work on one thing: giving AI reach into the physical world. Nanofabrication is our method. We are not building chips or a chatbot, and we are not working on DNA nanotechnology or biosecurity.
Xymos
Closing the gap between AI and the physical world.
Xymos is building a superintelligent AI system and connecting it to instruments that manipulate individual atoms. The aim is to give AI direct, general-purpose reach into the physical world.
01 Goal
AI is getting smarter faster than it is getting reach.
AI systems are becoming very good at reasoning, code and analysis. What they can change in the physical world is still limited by human-scale tools: labs, factories, robots, and the people who run them. Experiments are slow, expensive and mostly manual.
We think that, more than intelligence, is the ceiling on what AI does for the world.
We think the most general interface to the physical world is the atomic scale, because everything is built from atoms. A system that can place and measure them directly, and build better tools for doing so, is not limited to any one application. It could apply to medicine, materials, energy and security.
The thesis
The end goal is not nanofabrication. It is an AI that can act on the physical world directly, at the smallest scale matter allows.
02 Approach
Nanofabrication is how.
As far as anyone knows, physics permits atomically precise manufacturing. Feynman sketched the idea in 1959, and Drexler published detailed designs in 1992. In 1990, IBM researchers reported moving 35 xenon atoms one at a time to spell their company's name. It has not become general-purpose manufacturing at scale. We think the blocker is design and search, which is what a very capable AI is for.
- Step 1
Build a very capable system.
One general system, as capable as we can make it. We are not designing chips or new hardware.
Open problem - Step 2
Align it.
Make it do what we intend, and be able to check. This is the hardest open problem, and everything after it depends on it.
Unsolved - Step 3
Give it reach.
Connect it to instruments that place and measure single atoms, so it can run experiments on the physics directly instead of reasoning about it from papers.
Not started - Step 4
Close the loop.
Once it can fabricate at the nanoscale, each round of fabrication should improve the next, including copies of itself. That is the general-purpose interface.
Not started
03 Principles
How we think about the work.
Intelligence is necessary, not sufficient.
A capable model still needs instruments and feedback from real physics. Step three exists to close that gap, so the system learns from experiments and not only from papers.
Alignment is the hard part.
A system this capable, acting directly on matter and eventually copying itself, has to do what we intend. We don't yet know how to guarantee that, and no one else does either.
Say what we don't know.
We are early. We have no results and no proof of safety, and we will say so here until that changes.
04 Contact
Contact.
We're at the very beginning. If you work on alignment, nanofabrication or frontier AI, or want to support this, we'd like to hear from you.
hello@xymos.org