I think this viewpoint too much anthropocentric. Of course knowledge about riding a bike or doing a backflip can be accurately described in bullet points in ter

I think this viewpoint too much anthropocentric. Of course knowledge about riding a bike or doing a backflip can be accurately described in bullet points in terms of physics, it's just that people are not equipped to accurately follow those instructions, so we resort to trial and error. If a futuristic robot was given the bullet points to ride a bicycle, and it had full knowledge of its body, it should succeed on the first try.