AGI

AGI (Artificial General Intelligence) is the hypothetical milestone of an AI system that can understand, learn, and perform across essentially the full range of intellectual tasks a human can — and transfer knowledge between them — rather than being trained for one job. The contrast is narrow AI: today’s systems, including GPT-style large language models, are extraordinarily capable but still fundamentally task- and distribution-bound, not generally agentic minds. There is no agreed definition, threshold, or benchmark for AGI; proposals range from “matches a competent human at most economically valuable work” to “can do anything a human can,” and reasonable people disagree on whether current machine learning scaling is a path to it or a plateau short of it. A related term, ASI (Artificial Superintelligence), denotes a system that would exceed the best humans across the board. Treat “AGI” as a moving, contested goalpost and a marketing word as often as a technical one — when someone uses it, it’s worth asking what specific capability they actually mean.