AGI stands for Artificial General Intelligence: the hypothetical model as capable as a person at everything. Hypothetical is the operative word. It does not exist, it is not sitting in anyone's warehouse, and it is not on a product roadmap you can buy from.
The term has a real place in research, where people argue seriously about what general capability would even mean and how anyone would recognize it. But that is not where you will usually meet the word. You will meet it in a pitch, a keynote, or a meeting, doing atmospheric work: making the near future feel inevitable so that the thing being sold today feels urgent. When AGI is dropped into a business conversation, someone is usually selling something.
For an operator, the word is best treated as a filter. The moment a pitch leans on it, move the conversation back to ground level: what does the product do today, on our work, and what will it cost to run? Every real gain available to a business right now comes from the models that actually exist, applied to specific jobs with specific proof. None of it requires a machine as capable as a person at everything. It requires this quarter's problem, solved.
A concrete example. Two vendors pitch the same automation. One walks through your workflow and shows the system running on work like yours. The other talks about the coming age of general intelligence and the danger of being left behind. The first is selling you a tool. The second is selling you a feeling, and the difference between those two meetings is the entire practical content of the term.
“Let's solve this quarter's problem. AGI isn't going to ship it for us.”
// this page is the full text. the packaged PDF, all fifteen terms, comes with the free library.