It's a metaphor. Let me extend it a bit: Speech is the drought of repeated lies about the illegitimacy of an unwelcome outcome drying out the underbrush of widespread discontent.

Speech gathers the tinder of a crowd committed to "stop the steal".

Speech is the accelerant convincing those so predisposed that violence or the threat of violence is acceptable and even to be admired.

And speech is the match tossed offhandedly aiming the mob you primed and egged on at the target you want to intimidate (and that's the most charitable interpretation).

> you've placed the cause of this on algorithmic ranking as much as speech itself.

To be clear, I find algorithmic ranking to largely be the cause of mostly the 'widespread discontent based on repeated lies' part, although there are clearly also various 'automated radicalization' effects happening as well.

> Great, let's stop interfering with speech, it leads to bad things.

I don't think we get to unring that particular bell (especially since such ranking efforts predate both Twitter and Facebook). Applying ranking techniques to the prioritization and selection of a subset of items from a variety of sources is one of those ideas that was inevitable because it was obvious to a sufficiently large number of people, whether the result is a live feed, an automatically arranged 'front page', a playlist, or some other format. It doesn't even matter which specific techniques happened to be used, they were all going to be tried by a lot of folks until somebody got the results they wanted (more engagement / time on screen). Tech companies like Twitter and Facebook are going to continue to try to mitigate the toxic side effects on public discourse, but I don't think there is a way for them to back away from algorithmic ranking unless government regulation simply forbids it outright (which isn't too likely). For that matter, I'm not sure how useful HN itself would be if the front page wasn't being automatically ranked based on engagement+recency with editorial decisions being made to nuke certain types of posts/topics for being attractive nuisances.

Clearly, here's a better beginning:

“Let me extend the metaphor a bit".

> Speech is the drought of repeated lies about the illegitimacy of an unwelcome outcome drying out the underbrush of widespread discontent.

Did you mean hate speech? A drought is usually a negative. Sorry, I just can't make sense of that.

> And speech is the match tossed offhandedly aiming the mob you primed and egged on at the target you want to intimidate (and that's the most charitable interpretation).

All very poetic but, again, I'm sorry but I can't find any substance to it.

> I don't think we get to unring that particular bell

I wasn't the one suggesting we do. You presented two causes - algos and lies, the negative outcome of the latter being magnified by the former - and then chose to suppress one. I think that's a false choice as I can think of other possible ways to remedy the situation and would pursue them instead, but even if I only had the false choice then I wouldn't pick suppression of speech. I can manage with a chronological feed, I was on Twitter and Facebook pretty early, it wasn't bad enough for me to give up the foundation of liberal society. I even remember a time before the internet, it really wasn't that bad.

> I don't think there is a way for them to back away from algorithmic ranking unless government regulation simply forbids it outright (which isn't too likely)

What's likelihood got to do with it? You're just shrugging your shoulders and choosing to support the suppression of speech instead. It's not as if you've made the argument that you care about free speech but the alternatives are unlikely. I don't even see any evidence as to why it's unlikely. Making a decision based on the likelihood makes little sense, it's not a horse race.

> I'm not sure how useful HN itself would be if the front page wasn't being automatically ranked…

I agree. Why would legislation of algorithms mean no algorithms? It doesn't and it wouldn't. Here are some viable alternatives:

- Social media sites give users control over the algorithms that decide their feed - Users get access to the tools for blocking et al, and the scoring of users. Why can't I know that the person replying to me is likely to be a bot? Or has scored highly for trolling? Why can't I know my own score? Why can't I choose my own shadowbanning? No good reason other than the centralisation of power. - Companies over a certain size lose section 230 protections. This will encourage greater moderation on the large sites and foster competition from smaller ones.

I'm sure there are more.

Hiding some implementation details from bad actors isn't necessarily a bad idea, though of course you have to figure out whether you're doing the security equivalent of hiding a proprietary algorithm (bad idea) or hiding your secret key (good idea).