Things that support the notion that it is soon are that napkin math suggests the computational horsepower is here now, and that we have had few instances of sud

Things that support the notion that it is soon are that napkin math suggests the computational horsepower is here now, and that we have had few instances of sudden, unexpected advances in how well neural networks work. (Alpha Go, Alpha Zero, etc). One might extrapolate that there is a chance that in 10 years, when the computational horsepower is available to more researchers to play with, and we get another step-change advance, that we will get there. My own feeling is that it is possible AGI could happen soon, but I don't expect it will.