The theory is that we’re going to develop simulations of existing brains (inside computers) before we develop artificial intelligence. And, subsequently, we’ll developed incredibly cheap simulations of existing brains (inside incredibly cheap computers) at some point not terribly far after that.
So the “technological singularity” becomes one of trillions of brains able to solve problems, rather than, or in parallel to the development of strong artificial intelligence. In the podcast he calls these ’ems’ – i.e. emulations. The implication of this is that the value of labor plummets, while the value of raw materials (that this labor needs to be created) skyrockets.
The business/investing thought that follows is to invest in companies that produce and manage physical commodities, such as rare metals and energy needed to house and activate these ems. Another sensible investment (to me) is companies that produce computer technologies. For example, without Intel, ARM, IBM, Motorola, etc, there’s no hardware for these ems.