Does anyone seriously think that results of any current approaches would suddenly turn into godlike, super-intelligent AGI if only we threw an arbitrary number

Does anyone seriously think that results of any current approaches would suddenly turn into godlike, super-intelligent AGI if only we threw an arbitrary number of GPUs at them? I guess I assumed everyone believed this was a stepping stone at best, but were happy that it turned out to have some utility.