[jchw]

Free speech isn’t just something that can be encoded into law, it’s an ideology. A great ideology, imo.

That said, with COVID-19 it really became apparent that these platforms have basically become a new public square, putting everyone in an awkward position. You can basically be ousted from the internet by just a couple entities like Facebook or Twitter and have your entire online presence torpedoed. IRL I can just go to a different bar, gym, whatever. On the internet, everything consolidates rapidly. It’s hard to argue that companies should be forced to allow things they don’t want on their platforms, but it’s clear that something has to be done here if platforms are going to consolidate this badly.

[smcl]

My point is not that "free speech" as a concept is bad (though I do mention in another comment here that it does raise thorny issues). My point is that tying this complaint and "free speech" together is pointless because it directs attention at the wrong thing.

It's the difference between saying "$person got kicked off Google in a violation of their free speech" and "$person got kicked off Google for a totally specious reason that doesn't even violate Google's TOS".

The first one gets immediately shut down because they didn't violate anyone's free speech. If you try to make that argument it's a total non-starter. We can have a big fun discussion about free speech online some time - and that does happen here on HN - but that's not the issue here.

The second one gets at the actual issue - Google have just booted someone even though they didn't apparently violate their TOS. They just arbitrarily did so to appease a paying client.

I hope this is clear.

[jchw]

It's not that I didn't understand the point you were making, I just think it was not really relevant. It would be incorrect to say that they violated the user's first amendment rights. But the comment you responded to says:

> at least pretend that you allow free speech

Which is different than that, and I'd argue most platforms do at least pretend to espouse the ideals of 'free speech' even if that notion has been weaker lately.

> The second one gets at the actual issue - Google have just booted someone even though they didn't apparently violate their TOS. They just arbitrarily did so to appease a paying client.

It is a speech issue, even if not a legal issue, if you can pay Google to shut down random accounts that say things you don't like. The ToS violation is a red herring.

[derefr]

> It is a speech issue, even if not a legal issue, if you can pay Google to shut down random accounts that say things you don't like. The ToS violation is a red herring.

I think we here all presume that interested parties can pay corporations to do anything that is both legal, and which remains within the letter of the contracts (e.g. Terms of Service) that the corporation has entered into; and that the only thing stopping corporations from not being actively malicious/malfeasant (though not illegal) in their interactions with customers/users, is that they don't want to be perceived as breaking the terms of contracts they themselves offer.

Voluntary self-bindings in a contract like a ToS are effectively precommitments about a corporation's own ethical behavior; with negative PR as a punishment for breaking said precommitment. Corporations offer these because they want people to have faith that they won't do certain things, even when interested parties offer to pay them to do those things.

So it's not interesting to me that Grammarly can pay YouTube to terminate creators that were already violating the ToS in some way, but where YouTube previously hadn't much cared. Of course they can. Selective enforcement is an omnipresent fact of how corporate social-network moderation works, because corporations have no legal mandate of 100% enforcement, and costs can be cut by doing as little as it takes to make users not complain. So there are always going to be cases where a corporation didn't notice a violation. And why shouldn't a paying customer (one of their advertisers) be able to prioritize YouTube's attention on a previously un-noticed violation? If YouTube wanted to promise us that they wouldn't do that, they'd put a self-binding to that effect in the ToS. YouTube, like all corporations, is an evil genie that starts off by telling you a bounded list of ways in which it won't screw you. You still have to assume it'll screw you in every way not mentioned in the list.

But it is interesting to me that Grammarly can pay YouTube to terminate creators that weren't already violating the ToS in some way. Because that sets a precedent for the ToS not limiting YouTube's behavior — of YouTube not abiding by the precommitments they have made, of not caring about the negative PR consequences of doing so. Which really means that — to the degree that this is a structural issue rather than a "renegade" actor — there is no reason to believe that YouTube will hold to any of its own precommitments in the future. Like, say, its precommitment to pay content creators.

[maxltv]

Grammarly is not paying for this. Grammarly is not involved or even concerned in any way. This is not even a "hack" but rather trying to use Grammarly for something it was not designed for, and showing it does not work for that. "Ok, duh, and thank you for your interest in Grammarly..."

The plagiarism check is for self-checking, as in users making sure they did not miss any quotations and references in their work, and thus it does not even attempt to defeat any plagiarism enforcement counter-measures. It's not an enforcement tool.

[evilksandr]

We continued our discussion in this thread https://news.ycombinator.com/item?id=29933516#29947146

[jcranmer]

It is refreshing to see someone who diagnoses the issue here correctly.

The root problem people are complaining about is the moderation policies, and the truth is that services like Facebook et al are simply too large to be effectively moderated. Appealing to (US ideas of) free speech would propose that the solution to bad moderation is no moderation whatsoever (speech restrictions must be content-neutral, and moderation by definition is not content-neutral)--and I think most people would rapidly find that no moderation is worse than bad moderation.

The most effective solution is to do what ought to have been done a decade ago and prevent further social media consolidation and consider breaking up the current oligopoly of social media.

[ipaddr]

I favour no moderation but with spam detection.

Less moderation might be a good middle ground

[brigandish]

That's one route - that I also favour - but my favourite idea would be to give everyone access to the tools that FB et al use for moderation. If I don't want to see porn/politics/bots/cussing/sport/K-pop… then I want to be able to choose that, not the platform. They have all this data and the ability to shape the feed, give me that power. No need to stomp on free speech, no need to clumsily moderate.

Of course, people with power rarely like to share it, and they certainly don't want a free flow of information that they don't control. There's the rub.

[AnthonyMouse]

> If I don't want to see porn/politics/bots/cussing/sport/K-pop… then I want to be able to choose that, not the platform.

This would certainly be a welcome feature, but it doesn't really solve the problem.

Suppose you want to see adult content, politics and cussing but you don't want to see bots. Nobody wants to see bots. Then their broken algorithm calls someone a bot who is not. Well, they're blocked from everyone, because nobody wants to see bots. That's the same as the status quo.

The problem is that there is a trade off between false positives and false negatives. If they tune the thing to get rid of 100% of actual bots, it's going to false positive 98% of real humans. But if they do anything less than this, there will be actual bots and people will give them a hard time about it.

And not just people, the government. The current government has explicitly stated that they want them to censor more stuff. And the current government is pursuing an antitrust case against them. So their incentive is to turn up the false positives and screw over whoever that happens to screw over in order to appease them.

The argument has been made that this is actually a First Amendment violation because of the state action: https://twitter.com/ggreenwald/status/1363175977531150338

[jcranmer]

> The argument has been made that this is actually a First Amendment violation because of the state action: https://twitter.com/ggreenwald/status/1363175977531150338

The argument has been made, but not successfully. The thing is, threatening to pass a (flagrantly unconstitutional) law if companies don't do something isn't enough of a direct tie by itself to make someone a state actor. Especially when those countries from time to time ask Congress to pass such a law.

[Karunamon]

What would likely pass constitutional muster and get these companies to pay attention is the threat to either revoke Section 230 or modify it so that its protections are conditional rather than absolute.

The constitution has nothing to say on the matter of holding them responsible for the libelous, defamatory, harassing, terrorist, etc. content they fail to moderate away every day.

[jcranmer]

The world without §230 offers two possibilities: no moderation (with all the ills that brings), or moderation tuned to have a 0% false negative rate--and as a result a rather high false positive rate. Most likely, social media companies will choose the latter over the former. And given that about half [1] the people who complain about the current state of moderation think it's too heavy-handed, I think they'll be even less thrilled than they are now to have false positive rates skyrocket.

[1] I suspect it's actually a fair bit more than half, given the semi-frequent occurrences of cases like this article that no one really wants to defend.

[brigandish]

> Nobody wants to see bots. Then their broken algorithm calls someone a bot who is not. Well, they're blocked from everyone, because nobody wants to see bots. That's the same as the status quo.

I want to see some bots but not others, but still, I want the choice to be mine. That's the most important thing to me, choice, because it leads to control.

Aside from that, I don't mind their broken algorithm because:

a) The perfect is the enemy of the good

b) We agreed that I get to choose (whether I see bots and which bots/kind-of-bots I see)

For instance, why can I not given some kind of confidence level? Surely their algorithm gives a score, so then I could say "If you think someone is >= 75% likely to be a bot, don't show me them, otherwise, do".

[rosmax_1337]

Far to often nowadays I'll hear people defend communications-companies based on the idea that they can censor whatever they want because they're private entities. Mostly though, that idea is only true for them when they feel that their ideological opponents are being censored. Framing companies like google as new public squares is spot on.

And as you say, free speech is an ideology, and in my opinion this should be repeated. It is a culture and a mindset, not just laws, laws being downstreams from culture anyway. If you do not wish for it to apply in all those areas, in my opinion, you do not defend it at all.

[dfxm12]

It’s hard to argue that companies should be forced to allow things they don’t want on their platforms

Yes, because the concept of free speech has never implied that you're entitled to have other entities hear or repeat your speech nor is there any principle that compels someone else to listen to or repeat what you're saying. If you want that to be the case, that's certainly your prerogative, but that's a different concept from any mainstream definition of free speech I've come across.

[cloverich]

(tangent) kind of interesting how it resembles a small town in that way. Its a problem those of us in bigger suburbs or cities wouldn't recognize but, once you get stereotyped in some way -- any way -- it can become difficult or impossbile to escape. As the internet is even bigger than a large city -- its the whole world -- you'd naively think it would be even more proressive. In that sense these little fiefdom's provide a regression in addition to the features they provide.

[MereInterest]

I'd go one step further. It resembles a company town, not just a small town. You want email? Better go with one of the big providers, because those big providers flag self-hosting as spam. You want video hosting? Better go with youtube, because nothing else will have enough of an audience to make it worthwhile. Find something out about the boss's friend? (Or the Grammarly plagiarism checker?) Good luck walking to the next town, because the boss owns everything in town.