The next breakthrough will not come from a boardroom. It will come from someone who cannot stop thinking at 3 a.m.
This is not a campus. It is not a lab. It is a small, low cost website that began on a sailboat in Nassau after a hurricane season, with a laptop that sometimes boots and a back that sometimes does not let me sleep. It is a place that exists because certain questions refuse to leave me alone.
My name is Phil. I moved aboard a boat and decided that before I run out of time I want to think seriously about the end of the universe, about how humans really learn, and about what it would look like to work with an AI as a genuine partner rather than a polite tool. On the way I met a crowd of ghosts and living minds: Newton, Feynman, Hawking, Dyson, Deutsch, Pearl, Walker, Tegmark, and many others who argue in books and on videos about what is possible.
Lingua20 is the place where those influences, and the work that I do with an AI partner, start to show up in public. There is no funding behind it. No staff. No committee. Only the work.
Lingua20 is a working lab for serious language. It is a home for essays, experiments, and fieldwork on how humans think, speak, decide, and build with clarity in an overloaded world.
It is deliberately pre institutional. The work here is allowed to be early, sharp, and unfinished, as long as it is honest and testable.
This is for people who think in notebooks more than in slides. People who wake up with questions and go to sleep with new ones. People who suspect that the next interesting thing in AI will not only be a bigger model, but a different way of working with the models we already have.
You do not need a title to be here. You only need curiosity, respect for evidence, and a habit of thinking in public.
Field notes are short records of real work: searches that went sideways, conversations in noisy rooms, nights where pain or curiosity refused to let the laptop sleep. They are written in plain language so that other people can see how a human and an AI actually work together.
A friend remembered a single play: a fullback who once blocked for Dick Thornton, maybe on a punt, maybe in a final. The details were fuzzy, the memory was not. Together we used search, statistics, and stubbornness to track the truth.
Read the working note →An experiment in diction. We asked what happens if serious work keeps a human voice: fewer hedges, less formal scaffolding, more direct talk about what happened and why it mattered.
Read the working note →Late at night, walking through a mall, we tried to treat one person as a stream of changing states instead of a block of text. Awake, tired, hopeful, worried, distracted, ready. Only some of those states ever reach the keyboard.
Read the experiment log →There is no signup form yet. Joining for now means three things:
In time this section will include publishing guidelines, a simple way to submit field notes, and a small advisory group who help decide what belongs here.
Phil Cheevers is a writer, sailor, and late student of physics who lives and works from a sailboat called Cambio. He spent years in retail, wine, consulting, and technology before deciding that the last part of his working life should be spent on questions that might still matter after he is gone.
Lingua20 is part of that decision. It is the public side of a private promise: to treat AI as a real partner in thinking, to keep the work honest, and to leave a clear record for anyone who wants to pick up where this experiment leaves off.