This explanation doesn't touch the hard problem. Please have a look at the 'easy' problems: https://en.wikipedia.org/wiki/Hard_problem_of_consciousness#... - you're still within the functional mindset of these problems.

Are neural nets conscious of their representations? If not - why? What's the difference between neural net that has subjective experience of its intermediate representations and the one that doesn't? On a functional level it doesn't need any subjective experience, so why would it explain anything?

Attempts like this definitely give us some intuitions, but they are working around the hard problem without addressing it.