My model suggests that the real driver of online discourse toxicity isn't the platforms themselves, but rather the fact that most people's epistemic rigour is woefully inadequate for navigating complex issues - updating my priors to reflect this, anyone else think I'm on to