I wonder how long until techniques like Depth Anything (https://depth-anything-v2.github.io/ https://depth-anything-v2.github.io/) provide parity with human dep

I wonder how long until techniques like Depth Anything (https://depth-anything-v2.github.io/ https://depth-anything-v2.github.io/) provide parity with human depth perception. In Mark Rober's tests, I'm not sure even a human would have passed the fog scenario, however.