In reality, when hate speech speakers congregate, they eventually end up being forced to stop by an ever-growing collective of people who will not tolerate hate speech; with speech, courts, laws, and violence.

In the world you describe with telepathy, we would know instantly who was thinking hate speech, and be able to identify them and protect those they hate from them, and prevent them from congregating into cancerous hate mobs. Some would argue that this is thoughtcrime enforcement, which is an expected outcome from granting telepathy to all. That is, in a nutshell, correct: peer enforcement of group morality in a telepathic society would, by definition, put the “thought” in “thoughtcrime”.

However, in today’s non-telepathic world, we cannot easily identify those who use hate speech online and protect others from their hate. This is presenting many difficulties not addressed by your telepathic example, and is the core of the problem that must be addressed.