VIVI Tech
AI that never stops learning.
“A brain runs on a cup of coffee and keeps its own secrets. Why does AI need a power plant and all of yours?”
A brain fits a lifetime in a grapefruit, and learns new tricks without forgetting the old. Models do neither; VIVI Tech does both. You decide what it keeps.
Data storage for AI.
It started as a question: “Why couldn’t we store data as memories?” This filing is the answer, a dataset’s meaning packed into a compact VIVI card a model learns from. And the card is yours: add to it, replace parts of it, repair it.
Continuous learning.
Today, when a model must learn one new thing, you retrain the whole thing. The film is about that bill. This filing is the other way: new knowledge without overwriting the old, and forgetting on request. What it forgets is your call.
The model stays current. Two patents filed.
And the two stack: the model keeps learning against the card itself.
The Dendritron.
Richard Aragon’s research on a neuron whose branches learn locally, so new learning stays out of old learning’s way. It is early, and it is published in the open. Research, not a product, and not one of the filings above.
How the math actually works.
The pictures above are the plain-language version. The deep one is published in the open at our research arm — the papers, the code, the running work. If you want to go hard, this is the door.