There may be philosophical (i.e. fundamental) challenges to AGI. Consider, e.g., Godel's Incompleteness Theorem. Though Scott Aaronson argues this does not matter (see e.g., youtube video, "How Much Math Is Knowable?"). There would also seem to be limits to the computation of potentially chaotic systems. And in general, verifying physical theories has required the carrying out of actual physical experiment. Even if we were to build a fully reasoning model, "pondering" is not always sufficient.
It’s also easy to forget that “reason is the slave of the passions” (Hume) - a lot of what we regard as intelligence is explicitly tied to other, baser (or more elevated) parts of the human experience.
of course we can have AGI (damned if we don't) because we put so much, it better works
but the problem we cant do that right because its so expensive, AGI is not matter of if but when
but even then it always about the cost