No—we will achieve neither, anytime soon.

Conditioned on human survival, yes. Without such a condition, no.

Increasing the generality of AI has been repeatedly demonstrated to be possible, and it comes with major benefits/advantages. These advantages if pursued are likely to soon include the ability for AI to design its own specialized/narrow AI tools without human help and above a human level, in which case the economic value of a human AI researcher would be less than zero. At some point, these accumulated advantages will most likely cause humanity to be disempowered and discarded. That's pretty transformative, but probably not what Will was talking about.

If you want the mundane "hockey-stick economy" kind of transformative rather than the insane-but-default "superintelligence" kind, you have to stay alive long enough to appreciate it, which involves not creating strong AGI / ASI. In a Pause scenario, I foresee a lot of investment in fairly transformative narrow AI tools.