That’s a good start, but doesn’t help us since even the most basic programs can fit that definition and are not what we mean when we say AGI. I have yet to see anything approaching a useful definition of intelligence across several discussions of impressive new language, and other statistical models - it feels like we should have that before talking about making it real!
With fire, even “when things turn from themselves to ashes and produce heat” (which I imagine a prehistoric child could come up with) distinguishes fire usefully from most other phenomena in the world
You are equivocating throughout this thread: you accept that we developed an understanding of fire incrementally, and concurrently developed fire-based technologies, yet you insist it is different for intelligence, without giving any good reason to think so.
With fire, even “when things turn from themselves to ashes and produce heat” (which I imagine a prehistoric child could come up with) distinguishes fire usefully from most other phenomena in the world