The bet is that when US labor participation rate is less than 10% then em-like automation will contribute more to GDP than AGI-like.
Robin Hanson wrote the book the Age of EM. The "Em" are virtual human minds with no physical body or very tiny physical bodies. Futurologist Robin Hanson works George Mason University as a Professor of Finance and Cato Adjunct Scholar John H. Cochrane.
Hanson believes that scanning the brain in fine detail and then emulating the brain in hardware is the faster path to artificial human or artificial general intelligence.
I, Brian Wang, do not see a definition of AGI-like AI in the bet/prediction. I presume they mean all other paths to advance AI other than whole brain emulation or partial brain emulation or some kind of bootstrapping intelligence leverage biological copying and perhaps brain computer interfaces.