This isn't really any different from a "too big to fail"-kind of argument. We shouldn't hold websites liable for the algorithms they develop in order to generat

This isn't really any different from a "too big to fail"-kind of argument. We shouldn't hold websites liable for the algorithms they develop in order to generate profits, because it would be so inconvenient for them as to ruin their business model? That doesn't strike me as a good argument, and I have less and less of a personal belief that section 230 is a good thing. I think it's certainly the reason that we have the large internet companies we do today. They get to externalize costs in a way no other companies get to, by not being liable for harms associated with their own business models. Rampant copyright infringement, we'll just let you make ContentID as a compromise. Rampant child pornography and child exploitation, we'll just have you do the same thing but with CP data sets... No actual physical business would be able to get away with business models that presented such realities, but we give carte blanche to the internet companies to do exactly that.