[hybridity]

When I give ChatGPT a piece of literature or code that it has never seen before, it is able to analyze it and respond back at least semi-intelligibly.

This would make me conclude that the models we have today can in fact read, and typically rather well.

[humanistbot]

That is because it has seen literature or code that is syntactically similar enough to what you give it, so that it can autocomplete its way to a semi-intelligent response.