Our commitment to free speech is tested by speech that offends us, not by speech we agree with. This does not necessarily entail allowing all speech: it’s possible to take the pragmatic view that while we’d ideally want all speech to be permissible, it might be the case that in some contexts, the risks of violence (or other negative consequences) are too great.
I’m not going to repeat the standard arguments in favour of freedom of speech here (previous defences of the principle can be found in this column on Kuli Roberts, this one on Gareth Cliff, or this one on more general issues to do with “thoughtcrime” and hate speech).
On this pragmatic reasoning, one might ask how we most efficiently nudge ourselves into a world where all speech is allowed, even as those who utter hateful speech pay some other price (for example, widespread opprobrium) for doing so?