Apple's and oranges. Harassment is a specific legally definable thing that we have legal processes in place to handle. I don't see why platforms can't model their harassment policies off of that, and enforce them impartially, or just have the courts handle it.

Hate speech is not something you can nail down, because it's not a real thing. The working definition of hate speech, as it's actually used is "Anything I don't like". I could very well say that you are being hateful against right wingers (I'm not saying that, but I COULD), and it would be about as sound an argument as any other for hate speech.

So yes, if a platform has a monopoly, then they should be coerced into allowing "hate speech".

EDIT: I should probably make it clear that my preferred solution is to force (yes, force) Twitter and Facebook to make themselves compatible with API's like activitypub or similar. I understand that's a pandora's box, but we seem to be in Pandora's emporium in this situation, and must open some box or other.