Learning in a simulator

See all posts

The simulator doesn’t have to be perfect because the adaptation process which deals for variation from one world condition to another also deals with the variation from the simulator to the real world (33:49)

One of Talik’s takes is that starting with language may be less ideal becuase it differs from the proven bio-stack. That starting with language doesn’t fit in with the rest of biology. However, I think babies are more intelligent than we give them credit for. Humans naturally start exposing newborns to language (“mama” / “papa”) very early on, often before they develop motor skills (walking/crawling/even eating). To be fair, innate motor skills such as suckling and breathing already exist. But are innate motor skills necessary?

Minimal human API?