Language models were built to write. Their smallest unit is a word.
SLX is an *intent execution engine* — a different kind of AI, whose smallest unit is an action. Something that actually happens.
Most of what you'd want AI to do isn't a question. It's a job. And jobs are messy — the thing you need is in an app you're not in, or a file you can't find, or somebody else's inbox. That's not a flaw to be engineered away. It's what work is.
So SLX works out what you're actually trying to get done, finds where each piece of it lives, asks whoever has it, keeps what's private private, and carries the whole thing through to finished. Then it tells you exactly what it did.
It works on its own, for one person. It works better when there are other people in it. Same engine — it just has more to work with.
