An independent
laboratory.
Mycellium Lab is an independent scientific research, technology development, and project acceleration laboratory based in Florida, United States.
Research that stays
in contact with reality.
Mycellium exists to investigate how artificial systems remember, construct worlds, form agency, and learn when to act, defer, or wait, and to do it against real consequences rather than in isolation.
Artificial intelligence is the laboratory’s central instrument: a scientific instrument, an engineering medium, and an acceleration layer for research, discovery, simulation, decision support, and the creation of new systems.
Genesis is the research horizon. It is a question the laboratory is converting into testable hypotheses, not a capability it claims to have demonstrated.
Independent projects,
one shared core.
The laboratory is financed through the construction of real products. Research is not separated from engineering; each project is both a working system and a point of contact with reality.
Every project shares one engineering core of components, infrastructure, and method, while remaining an independent system with its own domain, users, and standards of evidence.
Projects exchange evidence with one another. What one system learns about uncertainty, provenance, or restraint can inform another, provided the finding survives verification in its own domain first.
A project may evolve into a product, a licensing arrangement, or an independent structure of its own. That is a normal outcome, not a departure from the laboratory.
OpenAI, Anthropic, and xAI for reasoning, extraction, and synthesis. Local models where latency or privacy requires it.
PyTorch and JAX for world-model research. GPU capacity on AWS, reserved rather than always on.
Cloudflare Workers at the edge for public surfaces. Hetzner for the long-running services and databases the lab depends on.
Postgres and SQLite as the record. Object storage for media. Every claim traceable to the source it came from.
How we decide
what not to do.
The laboratory holds a written ethos: scientific rigor that examines its own assumptions, artificial intelligence treated as an instrument rather than a destination, and ethics constructed in the details of daily work rather than reviewed after the fact.