I disagree with a lot if not all of this, but most frustrating was that I only see a single sentence in this that directly addresses the title, and it doesn't really elaborate on its hypothetical.

The overall story is also kind of a strawman. Anthropic doesn't really need to believe any of their Claude checkpoints is conscious and/or experiences qualia, they only need to find it possible that they might (one day).

That is to say nothing about how tormenting these models in itself is a form of anthropomorphization and parasocial behavior, and when it isn't, it's basically spam.