comments – Techdirt (original) (raw)

Techdirt Passes Two Million Comments!

from the seems-like-a-lot dept

Just a quick post to note an amazing (to me!) milestone. At some point last week (on Wednesday basically), this site passed over two million comments. That is since the site’s commenting feature launched in 1999. If you want the quick history: we started a newsletter 27 years ago on August 23, 1997, and it became a website in the spring of 1998, but we didn’t shift to the blog format with comments until March of 1999.

So, that’s basically two million comments across 25 years. Holy shit, that’s a lot of comments. Thank you to all of you who participate, especially the ones who add value with thoughtful, insightful, and funny comments. That’s what we’re always looking for.

It’s kind of incredible to me that this has lasted this long, especially given just how much the web has changed over this time, including how commenting has changed. People don’t remember this at all, but when Techdirt launched in the blog format (using Slashcode 0.3), it posted the email address publicly if a user entered their email address in the form. Because in those days, that’s what people expected. The idea that people might want to keep their email addresses private, or that spam would be a problem, wasn’t even part of the thought process!

How far we’ve come.

I had thought about figuring out which comment was the actual two millionth, but that’s complicated by lots of factors, including that there is still plenty of comment spam that we miss. Just a few days before we hit the two million comment mark, I happened across an article from years ago that had about 50 comment spam messages that we had missed at the time, but which I promptly deleted. So what was the actual two millionth comment isn’t really definable, as I could very well find another cache of old spam on another day and delete them as well.

And, of course, we get somewhere on the order of 5,000 attempts at comment spam a day which are blocked before they ever get on the site. Only a very small percentage of spam gets through (though it’s still frustrating). If we were counting the number of attempted comments, including spam, then we’d be many millions higher.

Still, thank you to the community here of (mostly) productive commenters who keep things interesting and keep us on our toes here. The community aspects of this site are always what make it the best.

Filed Under: comments, techdirt
Companies: techdirt

Techdirt 2023: The Stats

from the closing-the-books dept

Every year a little after New Years, I do a post about the previous year of Techdirt traffic and comments, looking at what people were interested in, what commenters were highly rated, etc. I know most sites put this out towards the end of the year, but I remain a purist and wait until after the new year begins to get all the stats. I usually try to do it a few days after New Year’s, but it’s a long process, and this year I’ve had a very busy start to the year (though, looking back, last year’s also came out about two weeks after the new year started as well).

If you’d like to see the details from previous years, here they are: 2022, 2020, 2019, 2018, 2017, 2016, 2015, 2014, 2013, 2012, 2011 and 2010.

Two years ago was the first year we did it without Google Analytics (which we ditched as we tried to remove as much Google as possible from the site). And last year it was a bit trickier, because in early 2022 we had switched from our old, homemade platform to WordPress, and that mucked up some of the stat tracking, and left me trying to piece things together.

This time we have a full year of data again, coming from two separate tools: Automattic’s JetPack and Plausible, which provides very simple, privacy-protective analytics data without feeding it into an advertising juggernaut.

As is pretty typical, about 72% of our traffic comes from the US, followed by Canada, the UK, Australia, Germany, India, the Netherlands, France, New Zealand, and Sweden. The differences from last year are marginal, but Germany was behind India last year, and Brazil was in the top ten (at spot 10) while New Zealand was not. Brazil was 15th this year, trailing Ireland, Spain, Italy and Switzerland. The top Asian country was Japan (one spot behind Brazil) followed by Singapore (though Japan and Singapore had nearly identical traffic).

In 2023 we published 2,007 posts (a tiny bit down from last year, but close) and garnered 61,056 comments on those posts, an average of 30.4 comments per post. Given commenting rates, it’s likely that this year we’ll pass 2 million total comments (!!!) somewhere in the summertime.

Of course, even as we published fewer posts, they got longer. Our average words per post was 798, up from 774 the year before. So even as we published fewer overall posts, we published more words. Just for fun, check out this chart:

Already, for 2024, we’re averaging (well) over 800 words per post, though who knows if that will stick. But, still, most people don’t realize for the first few years I though all Techdirt posts should only be one paragraph. That… changed.

In terms of traffic referrals, I always highlight that, by far, the largest source of our traffic is direct traffic. We never played social media games, trying to goose our traffic that way. And while that maybe put us behind others in overall traffic, our readers tend to be more loyal, rather than drive-by. It looks like approximately 50% of our traffic was direct, with no referrals. Google search sent about 17% of our traffic, and Google News drove about 6%.

After that, we did still receive about 5% of our traffic from ExTwitter even though we don’t post there any more (either personally or from the Techdirt account, since Elon took away the API to do so). Reddit and Smartnews each drove about the same level of traffic as ExTwitter. I honestly still don’t know what Smartnews is, but every year it sends a decent amount of traffic. Same with the “NewsBreakApp.” No idea, but thanks for the traffic.

This year, Flipboard also sent a decent amount of traffic, which is cool, given how that company is embracing the fediverse (something I’m hoping to write more about soon). After that we had HackerNews, Bluesky (my main daily social media app, which is getting close to fully opening to the public) and even Substack. Fark and LinkedIn both also sent a surprising amount of traffic.

Traffic from other publications had Ars Technica leading the way, followed by The Verge, Naked Capitalism, Daring Fireball, Kottke, Techmeme, and AboveTheLaw. I guess Techmeme is more of an aggregator, but it feels appropriate here. I like to see some of these oldschool blogs (Daring Fireball! Kottke!) in the list.

In search, after Google, DuckDuckGo and Bing were next in the list, but there were large periods of last year where Techdirt was missing from both Bing and DuckDuckGo (it looks like we’re currently back).

It’s been kinda crazy to watch the transition to mobile over the years (and we long resisted having a mobile-friendly site). But in 2023, it looks like 69.3% of Techdirt’s traffic was from mobile devices (phones or tablets), and just 30.7% from computers (desktop or laptop).

In terms of OS, 37.7% were Android, 31.6% were iOS. 18.1% were Windows. 10.7% were Mac. Linux was at 1.6% (though I just set up my laptop to dual boot into Linux, so we’ll see if I can bump that number up this year). ChromeOS rounds it out at just 0.5%.

Okay, onto the lists!

Top Ten Stories, by unique pageviews, on Techdirt for 2023:

I’d say there’s a good mix of expected ones and surprises in there. It does seem like “companies behaving badly” often gets a fair bit of attention from readers…

2023’s Top Ten Stories, by comment volume:

Also noticing a bit of a pattern here (and you might too if you went into the comments). We sure do have some extremely committed commenters.

And… once again, as we point out almost every year, there’s no overlap between the highest trafficked posts and the posts with the most comments, even if there’s a common theme in both lists.

Now, to the personal commenter leaderboards:

2023 Top Commenters, by comment volume:

Stephen T. Stone 3492 comments
That One Guy 1721 comments
Toom1275 1678 comments
bhull242 1281 comments
PaulT 1086 comments
LostInLoDOS 922 comments
Strawb 727 comments
Samuel Abram 706 comments
Mike Masnick 524 comments
Anathema Device 508 comments

Some expected names on that list and a few new ones as well. Second year in a row that Stephen T. Stone was atop the list, though he’s been hovering around the top 3 for years. But this is also the second year in a row that he posted nearly double the comments of the second place finisher, effectively looping the pack. Stephen, you could create another account, split your posts, and you’d still be in spots one and two…

Top 10 Most Insightful Commenters, based on how many times they got the lightbulb icon:

Parentheses shows what percentage of their comments got the icon

Stephen T. Stone 702 comments (20.1%)
That One Guy 497 comments (28.9%)
PaulT 199 comments (18.3%)
Mike Masnick 191 comments (36%)
Strawb 179 comments (24.6%)
Toom1275 141 comments (8.4%)
bhull242 139 comments (10.9%)
Thad 96 comments (25%)
JMT 83 comments (20.7%)
That Anonymous Coward 75 comments (14.9%)

Same top three as last year, and the same top three (in varying orders) as it has been for years. It’s great to have a crew of reliable, insightful commenters here.

Top 10 Funniest Commenters, based on how many times they got the laughing face icon:

Parentheses shows what percentage of their comments got the icon

Stephen T. Stone 124 comments (3.6%)
That One Guy 76 comments (4.4%)
Thad 42 comments (10.9%)
Strawb 34 comments (4.7%)
Toom1275 33 comments (2.0%)
Anathema Device 26 comments (5.1%)
Cat_Daddy 21 comments (11.3%)
That Anonymous Coward 21 comments (4.2%)
Samuel Abram 20 comments (2.8%)
Bloof 12 comments (6.6%)

Well done. As always, it’s harder to get the funny icon than the insightful one (perhaps we should fiddle with the thresholds?) But if you’re even remotely funny, it seems like it should be possible to get on this list next year.

Also, a shoutout to Thad for having a consistently high percentage of comments getting either insightful or funny, or both. Pretty impressive.

And, with that, the 2023 books are officially closed. 2024 is just a couple weeks in and I see that the competition is already pretty fierce for next year’s lists…

Filed Under: 2023, comments, stats, techdirt

The Patent Troll Lobby Set Up An AI-Powered Comment Creator To Support Its Bad Patent Policy

from the all-regulatory-comments-are-now-chatgpt's-opinion dept

You may recall that, back during the last net neutrality open comment period, the FCC’s comment system was overrun by millions of faked comments, including from many dead people. Not surprisingly, it was eventually determined that legacy broadband companies funded the fake comment submissions, which they felt they needed to do because actual activists were actually effective in getting the public to speak out in favor of net neutrality.

But of course, now we live in the “age of generative AI,” and it’s worth wondering just how that’s going to impact all of this. Amusingly, there are already academic journals suggesting that the government should sort and maybe even respond to regulatory open comments using AI as a tool.

But what about commenters themselves using AI to generate the comments in question?

We just recently wrote about how the US Patent Office is seeking comments on a dangerously problematic plan to make it much harder to kill bad patents by reforming the IPR (inter partes review) process to allow the patent director to just flat out reject challenges for certain classes of inventors, including many patent trolls. In that post, we linked to an EFF page urging people to send in their own comments against this proposal. But, really, the EFF is just linking people directly to the page on the Federal Register where you can comment. While they suggest some language, on the whole they expect users to write their own comment.

It appears the patent trolls who want this change have decided to use AI.

A few folks forwarded me copies of an email they received from “US Inventor” which is, effectively, a lobbying trade group for patent trolls, telling their members to submit a comment, and pointing them to an app on Streamlit. I’ll note that the email is ridiculous:

As you may know, the 2011 America Invents Act (AIA) created the Patent Trial and Appeal Board (PTAB). The PTAB is a nonjudicial administrative tribunal within the USPTO. The sole purpose of this court is to invalidate the same patents the USPTO previously granted, thus, creating a dictatorial power for the USPTO Director over America’s most important property right.

The statistics tell the truth about the destruction the PTAB has caused for inventors, small business owners, and startups. A staggering 69% of trials resulted in all claims being unpatentable, with an additional 15% leading to some claims being unpatentable. In total, 84% of patents reviewed by the PTAB result in the cancellation of claims.

I mean, no, as the Supreme Court itself made clear, just as the US PTO can grant a patent, so too can they review it, and then revoke it, if they realize they made a mistake. That’s not “dictatorial power.” And, let’s not even get started on the idea that patents are “America’s most important property right.” I mean, people with brains read this. Don’t insult them.

As for the second paragraph, a normal person would read that and think, wow, it sure looks like the IPR process that enables the PTAB to review patents is pretty damn important since it appears to be catching a very large number of mistakenly granted patents, which would otherwise gum up the work of actual innovation by blocking innovators from bringing products to market.

Because, remember, if the patent is a good, valid, patent, then the PTAB will not cancel claims. The only reason the PTAB cancels claims is if they are bad claims that should not have been granted in the first place, and would create an innovation-destroying monopoly power to block products that should be on the market. So even in this email, the troll lobby is effectively admitting that their real problem with the IPR/PTAB process is that it’s getting rid of their bad patent claims, that never should have been granted.

Anyway, the email sends people to a form that says will help you craft a comment. They don’t say it’s using AI, but it is.

You put in a link to a bio or a LinkedIn profile, some “information about you or your company” and then it generates a pro-patent troll argument for you. It gives you three attempts to do this.

The person who forwarded me the email also tried generating the letters, using the LinkedIn addresses of random people, and the “comments” it generated… were clearly just using AI to read someone’s LinkedIn bio to add some “pro-patent color” at the beginning, then some random AI-generated nonsense in the middle, before appending the identical text at the end of the comment.

Sometimes, the results are (in true generative AI fashion) total gibberish. One of the samples I saw definitely attacked the PTAB, but not for the IPR process. Instead, one claimed that the PTAB was changing claims to appear even broader than the inventor meant (which, um, is not happening).

Of course, the end of each of the “generated” comments is identical, and appears to be pre-filled in the comment generator app. No matter what information you put into the generator, at the end it pushes specific “policy alterations” that the submitter (who did not write the comment) claims should be considered.

Now, there’s an argument that making it easier to generate stronger comments during public comment periods might be a good thing in general. Having read some terrible comments submitted by the public, I could see some general value in letting people scribble in their general thoughts (like they normally submit) and having a tool turn it into something more substantive and useful. But… it also likely means that open comment systems for the federal government are even more likely to be overrun with questionable comments, probably many of which were not even generated by humans.

And that seems like it could be problematic for the overall process of regulatory bodies seeking public comments in the first place.

Filed Under: ai, comments, fake comments, ipr, patent trolls, patents, ptab, regulatory comments
Companies: us inventor

Techdirt 2022: The Stats

from the closing-the-books dept

Every year, a little after New Years, I try to do a post looking at the previous years results on Techdirt, what people were interested in, what commenters were rated highly and whatnot. I always wait until after New Years (unlike some other sites!) to make sure I have the full year’s data. This year, it took a little longer than usual as I’ve been pretty busy with some other stuff. Also, I had to do a bit more piecing of things together. As you may recall, last year, we finally switched from our old, home-built platform to WordPress. We’re still working out some of the bugs and quirks from the move, but slowly getting around to adding new features. But one of the issues is that we didn’t have a full, consistent analytics setup for the entire year, so I’m piecing things together from two separate analytics systems. As you’ll recall, two years ago we dumped Google Analytics in our ongoing quest to rely less on the biggest tech companies for services.

If you’d like to see the details from previous years, here you go: 2021, 2020, 2019, 2018, 2017, 2016, 2015, 2014, 2013, 2012, 2011, and 2010.

The first thing we cover is where our visitors came from… and the top of the list isn’t very surprising and has stayed pretty steady. 72% of our visits were from the US, with 6% from the UK and 5% from Canada. That’s in the range where we usually see it. Australia and India follow at 2% of our traffic. The next five are Germany, France, the Netherlands, Sweden and Brazil. I will note that one of the other analytics packages we’re testing (for which I only have a few months of data towards the end of the year) shows Brazil actually having the second most traffic, but I’m not sure that’s accurate.

I sometimes have fun looking through the bottom of the list, but honestly, that’s probably just a waste of time. The data on devices and browsers used to access Techdirt is a bit of a mess, though we should have much better data next year. However, it does look like the trend towards viewing via mobile devices continues to accelerate.

Some new stats that we didn’t have (easy) access to in previous years: in 2022, we published 2,056 posts and received just about 62,000 comments. Not bad! Those 2,056 posts totaled about 1.6 million words. That’s 774 words per post. Looking back at past data… that’s the longest average post we’ve had since the beginning of Techdirt. The data actually shows that our posts have tended to get longer each and every year with only a few exceptions. I honestly had no idea. I also don’t put much stake in overall traffic numbers, but our traffic increased a ton towards the end of last year. Basically from September onward, we were breaking traffic records.

The last few years I’ve given a handy pie chart for where our traffic comes from. I can’t quite do that this year (hopefully I can bring it back next year). Once again, we pride ourselves on having more than half of our traffic come direct (you loyal visitors coming right back, rather than relying on us popping up somewhere), and that still holds true. After that, search drove plenty of traffic, with (of course) Google driving most of that, followed by Bing and DuckDuckGo. Other traffic drivers were Twitter, SmartNews, Reddit, Google News, Facebook, and Hacker News. Yes, Facebook has always been a low traffic driver for us. We never spent much time cultivating traffic there like every other news site, and thus… we don’t much care how they handle news when every few months the company changes its mind.

And with that, let’s get to the stuff everyone looks forward to. The lists.

Top Ten Stories, by unique pageviews, on Techdirt for 2022:

This list is a bit different than in the past, in that much of it is kinda dominated by one story (and the next few posts down the list in traffic are related to that same story as well). That always makes me a bit nervous, as I prefer it when Techdirt is getting a pretty broad base of interest, rather than just focusing on a single thing. Still, the stories that did seem to catch on and go viral mostly tended to have our unique… Techdirtian spin on things. Which is something useful to think about for this year.

2022’s Top Ten Stories, by comment volume:

No One Has Any Clue How Texas’ Social Media Law Can Actually Work (Because It Can’t Work) (790 Comments)
5th Circuit Rewrites A Century Of 1st Amendment Law To Argue Internet Companies Have No Right To Moderate (625 Comments)
Stop Trying To Make State Action Doctrine Happen (467 Comments)
This Is Really, Really Dumb: Ohio Court Says Google May Be A Common Carrier (461 Comments)
The Problem With The Otherwise Very Good And Very Important Eleventh Circuit Decision On The Florida Social Media Law (452 Comments)
Twitter’s Legal Team Has Been An Aggressive Defender Of Free Speech; Will That Continue Under Musk? (420 Comments — Musk would find this funny)
No, The FBI Is NOT ‘Paying Twitter To Censor’ (409 Comments)
If You Think Free Speech Is Defined By Your Ability To Be An Asshole Without Consequence, You Don’t Understand Free Speech (But You Remain An Asshole) (393 Comments)
Danish Court Confirms Insane ‘Little Mermaid’ Copyright Ruling Against Newspaper Over Cartoon (387 Comments)
Musk, Twitter, Why The First Amendment Can’t Resolve Content Moderation (Part I) (351 Comments)

I am sensing a pattern here. People have strong opinions about the various laws trying to force websites to host content. And then some strong opinions about Twitter. I was somewhat surprised by the Little Mermaid story making the top 10 comment list, but then I looked and saw that our resident “very confused about copyright law” commenter went a little nuts on that one.

Also, despite the overlap in topics between these two top 10 lists, note that (once again) the top stories in traffic are not the same as the top stories in comments. Comments do not equal traffic. Sometimes they just equal flame wars between a small group of people.

And now to the really important lists. The comment leaderboards:

2022 Top Commenters, by comment volume:

Stephen T. Stone 3943 comments
That One Guy 2055 comments
LostInLoDOS 1985 comments
Toom1275 1923 comments
PaulT 1720 comments
That Anonymous Coward 1555 comments
Chozen 1385 comments
bhull242 1198 comments
Samuel Abram 951 comments
terop 914 comments

Definitely some carry-overs from the previous year, though PaulT, who has been somewhere on the leaderboard since we started this and was back in 1st place last year after a few years down the list, dropped back down to 5th place.

Top 10 Most Insightful Commenters, based on how many times they got the lightbulb icon:
Parentheses shows what percentage of their comments got the icon

Stephen T. Stone 668 comments (17%)
That One Guy 584 comments (28%)
PaulT 309 comments (18%)
That Anonymous Coward 180 comments (12%)
Toom1275 176 comments (9%)
Mike Masnick 147 comments (29%)
Thad 85 comments (21%)
James Burkhardt 70 comments (24%)
nasch 68 comments (8%)
Samuel Abram 67 comments (7%)

The same top three (though in a different order) as we’ve had for years. It’s perhaps not surprising Stephen T. Stone took first place this year, after his clean sweep of the most insightful comments of the year. Thanks guys for being truly key to providing real insight and value in the comments.

Top 10 Funniest Commenters, based on how many times they got the LOL icon:
Parentheses shows what percentage of their comments got the icon

That One Guy 87 comments (4.2%)
Stephen T. Stone 86 comments (2.2%)
Thad 40 comments (9.8%)
That Anonymous Coward 39 comments (2.5%)
Toom1275 36 comments (1.9%)
Cat_Daddy 26 comments (13.3%)
Samuel Abram 13 comments (1.4%)
PaulT 12 comments (0.7%)
Flakbait 11 comments (25.6%)
nasch 8 comments (1.0%)

Once again, it’s way more difficult to be consistently funny. Kudos to Thad who has stayed on this list for years with a consistently higher percentage than many others (and this year, much higher than normal). Also noteworthy are the two new entrants who both had an incredibly high percentage of their comments ranked as funny: Cat Daddy and (especially) Flakbait. Indeed, you have to go back to 2017 to find anyone with as high a percentage as those two (and, actually, Thad as well). Nice going guys! Keep the funny coming.

And, with that (a little later than usual) the 2022 books are closed and we’re off to the 2023 races. And some of you are already working hard on making next year’s lists…

Filed Under: 2022, comments, stats, techdirt

Court To Public University: Yeah, It’s A 1st Amendment Problem When You Delete Comments You Don’t Like

from the being-governed-by-Aggies dept

Just a somewhat periodic reminder: publicly-funded colleges are government entities. Almost every public university is. Sometimes, they seem to forget what they are and act in ways governments can’t — not without violating rights.

That inability to remember constraints imposed on it by inalienable rights is causing problems for Texas A&M University. And its arguments in court aren’t doing it any favors. A perennial protagonist has gone after the university in both regular courts and the court of public opinion. Its latter court appearances has prompted two civil rights lawsuits, detailed here by Eric Goldman.

Texas A&M (TAMU) does medical experiments on dogs. PETA objects to these experiments and commented on TAMU’s social media pages. TAMU blocked PETA, which led to a prior lawsuit that settled. The settlement terms included: “TAMU would not exercise viewpoint discrimination against PETA, its supporters, or members when administering its Facebook page; nor would it set automatic or manual blocking filters on PETA’s comments made to TAMU’s Facebook page, provided that TAMU could remove comments not in compliance with its Facebook Usage Policy.” Was there ever any doubt that this “resolution” would create further conflict?

Now, it’s well-known PETA courts controversy, routinely engages in highly performative anger, and otherwise often acts like a nuisance, rather than a public interest group. Every now and then, it manages to raise a valid issue.

And it’s not wrong here. And the settlement it obtained was never going to be enough to satisfy the demands of the Constitution.

A second lawsuit followed. PETA claims the settlement still allows the university to violate its First Amendment rights. The judge handling this case agrees with PETA, at least as far as procedural things are concerned. The lawsuit can move forward.

The judge is far less impressed with Texas A&M’s arguments in support of its motion to dismiss. Theoretically, A&M is the adult in the room. But you’d never be able to tell that from its terrible defense of ongoing censorship (real or theoretical) of PETA’s comments.

The court says [PDF] PETA has raised valid issues. This isn’t to say PETA will win. But it does definitely say Texas A&M can’t win — not with this settlement and not with these horrendous arguments.

The issue arises from comments PETA and its supporters added to a remote livestream of the school’s May 2020 graduation ceremony. PETA alleges the university [TAMU] deleted 137 of its 553 comments (most at Facebook, a few at YouTube).

The person representing the school is President M. Katherine Banks, named directly as the defendant in this follow-up suit. The school (through Banks) argues there’s no alleged injury, no ongoing injury, no potential future injury, and that the settlement made this all not worth arguing again in court.

First, Banks argued she shouldn’t even be named as a defendant, since it was her predecessor who crafted the inadequate settlement now being sued over. The court says she’s still on the hook. After all, lawsuits against other government agencies don’t simply end because the named defendant is no longer in office. The responsibility lies with the head of the government entity, whoever that currently happens to be.

Despite the inordinate number of pages devoted to this issue, which include a sur-reply, an advisory, and a reply to that advisory, I have no trouble finding that President Banks has a sufficient connection to the enforcement of TAMU’s social media policies. All that is required is that PETA allege enough facts that I can plausibly infer an enforcement connection. It is undisputed that President Young, TAMU’s previous President, executed a settlement agreement, in his official capacity, regarding the enforcement of TAMU’s social media policies against PETA. This action makes it entirely plausible that President Banks, as President Young’s successor, has at least “some connection” to the challenged behavior in this litigation. That is all that is required of this “straightforward inquiry.”

The court goes on to note that Banks provided plenty of precedent to support her argument, but none that actually supported her argument in this context. There is no precedent that agrees with the current university president’s argument she cannot be sued as the putative head of the entity engaging in the alleged violations.

The idea that a university’s head is an appropriate party for challenging the university’s actions seems so unobjectionable in a variety of contexts that I have not located cases holding otherwise, and President Banks certainly has not pointed me toward any.

The inability of TAMU’s president to find supporting precedent continues. Banks claimed no Constitutional injury in terms of future posting by PETA can be alleged because PETA failed to provide the school with details about its planned future actions in regards to TAMU social media content. Ridiculous, says the court. The allegation is enough to satisfy PETA’s standing requirements in this case. The arguments raised by Banks range from insipid to dangerous. If the university held off on deletion of PETA comments until after litigation concluded, but resumed shortly thereafter, what is even the point of having a court system?

[P]laying President Banks’s argument out to its logical conclusion shows how misguided it is. Were PETA to repost its original comments to TAMU’s social media sites and TAMU to delete them again, PETA would have the same standing it does now. But if—as is much more likely to occur given that litigation is ongoing—PETA were to repost its original comments to TAMU’s social media sites and TAMU did not delete them, I would nevertheless find that PETA has standing.

As the Supreme Court has repeatedly held:

It is well settled that a defendant’s voluntary cessation of a challenged practice does not deprive a federal court of its power to determine the legality of the practice. Such abandonment is an important factor bearing on the question whether a court should exercise its power to enjoin the defendant from renewing the practice, but that is a matter relating to the exercise rather than the existence of judicial power.

You can’t escape a lawsuit simply because you choose to not do the thing prompting the lawsuit temporarily. You cannot escape a lawsuit by making temporary concessions in hopes of eluding injury claims. Once you’re in it, you’re in it. Prove your case or face judgment.

PETA also claimed the limitations of the original settlement, along with TAMU’s refusal to engage outside of litigation, makes it unlikely PETA could be fully “restored” without the intervention of the court. President Banks claims otherwise. And, wow, is she wrong.

President Banks also argues that PETA’s restoration injury is not sufficiently imminent because it is premised on “three contingencies:

(1) individuals will watch TAMU’s over 500-day-old videos; (2) these viewers will read the comments posted to the videos; and (3) these viewers would have seen PETA’s deleted comments (if they were restored) while missing PETA’s many other non-deleted comments.” The only case President Banks cites in support of this metaphysical argument is Glass, 900 F.3d 233, a case as inapposite in this context as it was in the context of PETA’s posting injury.

A rights violation is still a violation whether or not only the entity suffering the violation notices it.

Just as a falling tree produces vibrational waves in the air, viewpoint discrimination offends the Constitution regardless of whether anyone is in the hypothetical forest to hear (or, in this case, read) it. Phrased differently, it does not matter whether anyone wants to read PETA’s comments; PETA has a right to put them in TAMU’s public forum.

How wrong can one university rep be? Very very very very wrong, apparently.

President Banks contends that PETA cannot assert organizational standing because it does not satisfy Article III’s standing requirements. But I have already determined that it does. That ought to be the end of the standing inquiry. However, President Banks argues that “to assert organizational standing, PETA must show that TAMU’s conduct ‘perceptibly impaired’ its mission.” Phrased differently, President Banks argues that it is not enough that PETA alleged a violation of its First Amendment right to free speech; PETA must also allege that the violation of its free speech rights perceptibly impaired its mission. This is obviously wrong.

Once again, cherry-picking rulings that seem to support your arguments is a losing strategy if the selective selection of cases don’t say the things you think they say.

Tellingly, none of the cases President Banks cites in support of this argument are cases regarding constitutional violations. The Constitution does not require this much. PETA “has alleged violations of its First Amendment . . . rights and thereby satisfied the irreparable injury requirement.”

The lawsuit moves forward and all claims, including PETA’s §1983 allegation, survive. TAMU is a government entity. No matter how much it may dislike PETA’s appearances in the court of public opinion, it cannot do what it did and expect to remain in constitutional good graces.

Filed Under: 1st amendment, comments, free speech
Companies: peta, texas a&m

China Unveils New Regulations Requiring Sites To Pre-Censor All Comments

from the yeah,-sure-that'll-work dept

As we see more and more western countries looking to regulate the internet in order to stifle speech they dislike, we’ve noted how much these efforts seem to be almost directly modeled on how China censors the internet. You might think that would be a reason to run in the other direction, but too many policymakers seem to now view China’s Great Firewall as a success story to be followed. And, now they may get some new ideas, as China has pushed out a draft of revisions to its regulations regarding online commenting. And, while some of it is unclear, it appears to include a provision saying that services that enable comments need to have tools in place to review every comment before it can be viewed on the site.

Specifically, the draft regulations include this section:

Establish and complete information security systems for the review and management, real-time inspection, emergency response, and the acceptance of reports for post comments, to review the content of post comments before publication, and promptly discover and address unlawful and negative information, and report it to the internet information departments.

For somewhat obvious reasons, that’s raising some concerns. As the Tech Review article linked above notes, online comments and other more real-time communications have always been a sort of loophole regarding the Great Firewall, as discussions on sensitive topics often breakthrough there, even if only to be deleted later. However, this new rule seems to be setting up a system to block even that.

There’s a need for a stand-alone regulation on comments because the vast number makes them difficult to censor as rigorously as other content, like articles or videos, says Eric Liu, a former censor for Weibo who’s now researching Chinese censorship at China Digital Times.

“One thing everyone in the censorship industry knows is that nobody pays attention to the replies and bullet chats. They are moderated carelessly, with minimum effort,” Liu says.

But recently, there have been several awkward cases where comments under government Weibo accounts went rogue, pointing out government lies or rejecting the official narrative. That could be what has prompted the regulator’s proposed update.

Tech Review quotes people saying that it’s unlikely (for now) that Beijing will require everyone to pre-review every comment (recognizing that’s likely to be impossible), but that it will put pressure on sites to be much more proactive, and that it could force this “feature” to be used on highly controversial topics.

It does seem that a straightforward reading of the law is that it requires sites to at least build out the functionality to pre-approve all comments if need be, even if it does not need to be on all the time.

There are some other features in the new regulations, including granting more power to who can block comments, suggesting that content creators themselves will have more power to censor comments in response to their content (rather than relying on the service’s in-house censors to do so).

Also, I note that part of these requirements would make Elon Musk and others who insist that every user should be “verified” even if their identities are not disclosed publicly, happy. As the rules require:

Follow the principle of ‘real names on file, but whatever you want up front’ , to conduct verification of identification information for registered users, and must not provide post comment services to users whose identification information has not been verified.

So, for all of the folks out there insisting that all internet users who are commenting should have identifying information on tap, in case it’s needed, just know that you’re following in the footsteps of Chinese censors.

And, of course, the new regulations also seek to tie that verified identity to China’s infamous social credit scoring system, though amusingly this is framed as part of privacy protections.

Establish and complete systems for the protection of users’ personal information: the handling of users’ personal information shall comply with the principles of legality, propriety, necessity, and creditworthiness; disclose rules for handling personal information: giving notice of the goals and methods of handling personal information, the types of personal information to be handled, the period for retention, and other such matters; and obtain the consent of the individuals in accordance with law, except as otherwise provided by laws and administrative regulations.

The people pushing for similar ideas in Europe and the US insist that it won’t be abused, but we can look to China — and the fact that many of the proposed regulations we’re seeing today originated as part of China’s Great Firewall for censorship to see where they likely lead.

Filed Under: censorship, china, comments, free speech, real names, social credit, verification

Techdirt 2021: The Stats.

from the closing-the-books dept

Every year, a few days after New Years (once the data is truly complete), I try to do a post exploring some of the traffic patterns and comment details on Techdirt for the year. This year’s will be a bit different on the traffic front, because at the end of last January, we took Google Analytics off the site, and that had been the tool we’d used to see where people were coming from and whatnot. Instead, this year for the details on where people are coming from and what technology they’re using, we’ll be using a combination of a self-hosted Matomo instance and Plausible Analytics (not self-hosted; though we’re hoping to eventually switch to the self-hosted version). We ran all three solutions for one month before turning off Google Analytics, and the data roughly, though not exactly, matched (because all traffic numbers are fake and unreliable), but this year’s info should be slightly different. As in the past, we’re not reporting silly things like how much traffic we get, because (again), all those numbers are fake, due to dodgy tools and even dodgier bots and such. But we can get useful comparative data about where people are coming from and such.

Even though it relied on different analytics, if you’d like to see the older versions, here they are: 2020, 2019, 2018, 2017, 2016, 2015, 2014, 2013, 2012, 2011 and 2010.

The first thing we usually cover is where visitors are coming from — and already we discover that with different tools, we get very different data. According to Plausible, our top five countries for visitors are the US at 70%, Canada at 5%, UK at 5%, Australia at 2% and India at 2%. Matomo has somewhat different results: US at 78%, UK at 6%, Canada at 3%, Australia at 2% and Germany at 1%. On Matomo, India actually comes in at number 15, rather than 5th on Plausible. Though Germany is 6th on Plausible. On Plausible the next four after Germany are France, Netherlands, Brazil and Sweden. On Matomo, the next five after Germany are France, Russia, Netherlands, Spain, and Sweden. Brazil, which came in 9th on Plausible shows up as 11th on Matomo. Russia, which shows up as 7th on Matomo is all the way down as 24th on Plausible!

I repeat: a lot of traffic numbers are garbage and no one knows anything! Last year we were surprised that Google Analytics said China was in our top 10, even though we had heard Techdirt was mostly blocked there. Matomo says that China gave us the 13th most visitors of any country, though Plausible says it’s actually 42nd (though, it counts Hong Kong separately, and has that come 37th — so perhaps if you combine the two…).

Moving on to what browsers people use, Chrome is still the leading tool for reading Techdirt. Matomo and Plausible seem to measure browser info in slightly different ways, but it does appear that somewhere around 45% of our traffic is from Chrome (Matomo breaks it down into different versions of Chrome) and somewhere between 25 and 32% from Safari. Firefox is somewhere between 8 and 12%. Interesting to note that most of the Safari traffic is from mobile (so, iOS) whereas most of the Chrome traffic is from a computer. There is still decent amount of mobile Chrome and desktop Safari, but it’s a lot more of the alternative. Microsoft Edge is around 4% according to both platforms. DuckDuckGo’s privacy browser is a new entrant this year, with both systems saying about 2% of our traffic is coming from there (which is cool to see). Both Opera and Samsung’s browser get about 1% each. Matomo claims that there were 126 different browsers used to visit Techdirt last year, which is quite something.

In terms of operating systems, Windows at 32% beat out iOS at 26%, Android at 24% and Mac at 14%. Linux variations were about 3% and Chrome OS was a bit under 1%. The two analytics systems seem to agree that about 48% of our traffic was from desktops/laptops, with 52% coming from phones/tablets. Incredibly, a tiny fraction of people apparently visited Techdirt from gaming consoles, TVs, or car browsers (?!?).

Every year I’m interested in, generally speaking, where visitors are coming from. While other sites spent lots of time “gaming” social media sites for traffic, we’ve always avoided that. I’m not against getting such traffic, but I don’t want to be reliant on it. So I’m always most interested in how much of our traffic is coming directly from people choosing to be here, and that continues to remain high, with 52% of our traffic coming from “direct” visits, 26.5% from search, only 14.2% from social media, and another 6.5% from other websites:

As for social media traffic, Twitter is the biggest single provider, with 55.6% of social media traffic coming from the bird site. Facebook provides 18.3% of the social media traffic, barely edging out Reddit by a barely noticeable number of visits. Hacker News comes in 4th at 6.3%. Everything beneath that is negligible, but if you must know, it’s LinkedIn, YouTube, and Telegram after the top four.

Google (84%) provides the majority of our search traffic (again, which is 26% of our overall traffic), followed by DuckDuckGo (10%), Bing (3%) and Yahoo (1%). If you mash social and search together, Google provides us the most traffic, followed by Twitter, Reddit, DuckDuckGo, then Facebook, HackerNews and Bing.

And… now we get to the lists.

Top Ten Stories, by unique pageviews, on Techdirt for 2021:

So… there are some oddities and surprises in that list. Four out of our top ten stories… are actually not from 2021! In fact, three out of the top four are not from 2021. That’s… weird! In the past, every so often we’ve had an older story slip into the top 10 list, but having four show up is very, very odd. In fact, the Parler story was also the number three story from 2020. It went a bit viral in January, when Parler got a lot of attention after being kicked off of Amazon and out of the various mobile app stores. The “Hello, You’ve Been Referred…” is not that surprising, as that’s taken on a life of its own and people now regularly link people to it on social media. These days, almost every time I come across someone spouting some nonsense about publisher v. platform, I’m relieved to see someone has already linked to my post. Service journalism.

The Kim Dotcom story is… from 2015. But it has snuck into the top 10 list in the past. In 2018, it also made a surprise appearance. Dotcom has had a pinned post on his Twitter linking to that story for years, so whenever he goes viral, that story tends to go viral as well.

Now, the oddest one of all: the 7th ranking story, the Neighbors are Watching Via Surveillance Video… is from 2003. And, honestly, until I was going through the stats to put together this list, I had no idea that it somehow had received a flood of attention. I don’t check stats that closely during the year, and am perplexed as to how or why this article from 18 years ago is suddenly in our top ten list. Here’s the other oddity: as far as I can tell, it’s not because of a bunch of traffic on a single day or two, but steady traffic from some sort of search (though it’s not clear what keyword is leading people to it) every damn day of the year. And not just Google. The search referral to that one story matches our general breakdown: Google searches supplied 88% of the traffic to that story, DuckDuckGo supplied 7%, and Bing supplied another 2%. On an average day it doesn’t get anywhere near the traffic of a new story, so I’d be unlikely to notice it. But it gets just enough traffic every single day that it’s in our top 10! I also see that comments continue to pour in on that story, which now has over 650 comments (though not that many in 2021). It appears that this has been a low-level traffic collector for a while now… I mean, in some ways, the story is interesting, in that it was a very early reporting on neighbor’s using surveillance cameras to help police — something that is now happening regularly via Ring and Nest cameras and the like. But I will admit that I am at a complete and total loss as to how or why that story gets so much traffic every day.

2021’s Top Ten Stories, by comment volume:

Huh. So the first few on that list are perhaps predictable, but then it goes off in directions I wouldn’t have expected. Once again, we see little overlap between the stories that get the most traffic and those that get the most comments, which always strikes me as worth noting (people always assume otherwise).

Okay, okay, I know what people are really waiting for are the comment leader boards…

2021 Top Commenters, by comment volume:

PaulT: 3419 comments
Stephen T. Stone: 2918 comments
Scary Devil Monastery: 2636 comments
That One Guy: 2334 comments
Lostinlodos: 1851 comments
That Anonymous Coward: 1301 comments
Toom1275: 1185 comments
Samuel Abram: 1172 comments
tp: 977 comments
ECA: 976 comments

The top four people are the same top four as last year, though in different order. PaulT returns to the top spot he last held in 2018. A couple of “new entrants” into the list this year, from two individuals who seem to spend more time arguing with the long time commenters, which seems to have contributed to their comment totals…

Top 10 Most Insightful Commenters, based on how many times they got the light bulb icon: Parentheses shows what percentage of their comments got the lightbulb

PaulT: 907 comments (26.5%)
That One Guy: 905 comments (38.8%)
Stephen T. Stone: 742 comments (25.4%)
Scary Devil Monastery: 254 comments (9.6%)
That Anonymous Coward: 246 comments (18.9%)
Toom1275: 232 comments (19.6%)
Mike Masnick: 201 comments (45.6%)
Bloof: 191 comments (43.12%)
Samuel Abram: 147 comments (12.5%)
Thad: 118 comments (26.5%)

Those top three really dominated the leadership boards again, as they’ve done for a few years now. Thank you for contributing such regularly insightful comments.

Top 10 Funniest Commenters, based on how many times they got the LOL icon: Parentheses shows what percentage of their comments got the LOL icon

Stephen T. Stone: 91 comments (3.1%)
That One Guy: 56 comments (2.4%)
That Anonymous Coward: 38 comments (2.9%)
Thad: 31 comments (7.0%)
Toom1275: 28 comments (2.4%)
Bloof: 25 comments (5.6%)
PaulT: 21 comments (0.6%)
Samuel Abram: 18 comments (1.5%)
Jojo 13 comments (12.4%)
Norahc: 11 comments (10.4%)

As always, we’ve seen that it’s much, much more difficult to get the coveted “funny” vote. Also, for the previous two years, we had noted that Norahc had come on strong as someone who commented less frequently than others, but had a really high percentage of comments voted as funny. This year, Norahc’s percentage of funny comments was even higher than the previous two years, but… newcomer Jojo stole the show by edging out Norahc with an even higher percentage of the funny vote.

And, with that, the 2021 books are closed, and 2022 is well underway. As noted in my final post of the year for 2021, we’ve got some decently big changes coming up soon, which should make the comments an even more fun place, so stay tuned…

Filed Under: 2021, comments, stats, techdirt

Wherein The Copia Institute Tells The Eleventh Circuit That Florida's SB 7072 Law Violates Our Rights

from the sticking-it-to-everyone-to-stick-it-to-facebook dept

We’ve talked a lot about the Florida law SB 7072 that attempts to regulate social media platforms. In broad strokes, it tries to constrain how at least certain Internet platforms moderate their platforms by imposing specific requirements on them about how they must or may not do so. That law is now being challenged in court. The district court enjoined it, and Florida has now appealed to the Eleventh Circuit to have the injunction overturned. This week the Copia Institute joined others in filing amicus briefs in support of maintaining the injunction.

As we told the told the court, the Copia Institute wears two hats: One hat we wear is as commentators on the issues raised by the intersection of technology and civil liberties, which laws like Florida’s impact. Meanwhile, the other hat is the one we wear by sitting at this crossroads ourselves, particularly with respect to free speech.

To operate Techdirt, the Copia Institute needs robust First Amendment protection, and also Section 230 protection, to both convey our own expression and to engage with our readers, including in our comments section. Unfortunately the Florida law impermissibly targets both sets of rights. And this constitutional and statutory incursion affects every Internet platform, and all the user speech they facilitate, including us and ours, even if we don’t all fall directly into its crosshairs.

The Florida law’s enforcement crosshairs are especially arbitrary, ostensibly targeting companies with very high revenue, or very large audiences, unless, of course, they happen to also own a theme park? But one thing we told the court is that the specific details don’t really bear on the law’s overall constitutional and statutory defects. Part of the reason is because if Florida could pick these arbitrary criteria, which might not apply to certain platforms, another state could pass a law with different criteria that would reach more, and then these platforms would still be left having to cope with a fundamentally impermissible law.

Also, it’s not clear that even small entities like ours might not be able to attract the larger audiences the Florida law describes since that’s at the very heart of what we try to do as an enterprise: have reach and influence. The point of the First Amendment is to make it possible for outlets like ours to connect with readers ? only thanks to laws like this, we could end up punished with onerous regulation we couldn’t possibly comply with should we succeed. And that sort of punitive deterrence to expression is not something the First Amendment, or even Section 230, permit.

But even if Techdirt could remain safe from the reach of a law like this, it would still hurt us if it hurt other platforms, because we need the help of other platforms to help our message get out too. Indeed, the whole point of the Florida law is ostensibly to help people use these other platforms to get their messages out. Only the upshot is that the law does the exact opposite by salting the regulatory earth so that no platform can safely exist to help users do that.

Filed Under: 11th circuit, amicus brief, comments, florida, free speech, moderation, section 230, social media

Killing Website Comment Sections Wasn't The Brilliant Move Many Newsroom Leaders Assumed

from the muzzles-aren't-innovation dept

Mon, Nov 8th 2021 09:30am - Karl Bode

So for years we pointed out how the trend of news websites killing off their comment section (usually because they were too cheap or lazy to creatively manage them) was counterproductive. One, it killed off a lot of local, community value and engagement created within your own properties. Two, it outsourced anything vaguely resembling functional conversation with your community — and a lot of additional impressions and engagement — to Facebook. Despite the downsides everybody ran with the idea that comment sections were utterly irredeemable and unnecessary.

Turns out, much of the conventional wisdom driving those decisions wasn’t so grounded in fact. This Poynter piece does a really good job revisiting whether killing the comment section was a good idea ten years on. It’s true that negative comments in the comment section can tarnish a visitor’s perception of the quality of an outlet’s brand. But it’s also true that the discussions outsourced to Facebook continue to also do that, they’re just doing that over at Facebook. So many researchers argue that if you’re going to have a discussion, you’re probably better off having at locally at your site:

“Conversations on Twitter, Facebook and Instagram won?t stop. And the same research premise holds true ? negative comments on those platforms will have a negative impact on the outlet?s credibility. So is it better to at least keep one forum where the outlet has control and the potential to monetize commenters into subscribers? And how do we make that forum as good as it can possibly be?”

If you recall, when most news empires over the last decade announced they were killing their comment sections, it was usually accompanied with some form of gibberish about how the decision was made because they just really “valued conversation” or wanted to “build better relationships.” Sometimes newsroom managers would be slightly more candid in acknowledging they just didn’t give enough of a shit to try very hard, in part because they felt news comments were just wild, untamable beasts, outside of the laws of physics and man, and irredeemable at best.

But again, as it turns out, none of that was true.

One recent experiment worked collaboratively with 24 Gannett newsrooms giving them four options: to turn off comments, to keep existing commenting systems in place, to use Vox Media?s “Coral” commenting system, but to use Coral?s commenting system and only allow subscribers to comment. You’ll never guess what the study found:

“Turning off comments actually lowered the average time readers spent on the site, according to Stroud?s research.

And journalists, who have the most to lose from a harsh comment, didn?t have increased job satisfaction or feel differently about how the newsroom served the community when comments were eliminated.”

While yes, many readers are often incoherent trolls, many other readers actually (gasp!) know what they’re talking about, and their input and conversations can actually improve journalism. As is often evident here at Techdirt, sometimes the resulting conversation can correct something the author has gotten wrong, or give reporters insights into trends and ideas they’d never previously even considered. If modern news is actually a conversation, quality comments are a helpful extension of that conversation:

The Detroit Free Press? Delgado sees involving reporters more routinely in the process as a potential solution. Having the journalist in the space with commenters can create a conversation between the newsroom and the community. It?s beneficial not just to readers, but to the reporters themselves.

?I know when I moderate comments, I?m a smarter, better journalist,? Delgado said. ?I know what people are talking about, and you can start to see a lot of the ideas and theories that are resonating.”

The problem wasn’t so much the comment section, it was poor managers running news organizations in a country that doesn’t properly fund journalism. And the study above does show that if you’re not going to run a comment section well, you’re better off not trying. But at the same time, a lot of these organizations did have the resources to do a better job at managing on-site community, it was just easier and cheaper to pretend comment sections were some irredeemable, malicious force we were all better off without to justify their corner cutting. That was always a narrow oversimplification.

The untapped irony is that many of these same major outlets that outsourced all discourse to Facebook over the last five to ten years, now complain incessantly about how Facebook has too much power over discourse, ad markets, and everything else. It’s pretty rare you’ll see anybody acknowledge that the decision to muzzle local communities and outsource all discourse to Facebook helped create at least some of the problems they’re now complaining about.

Filed Under: comments, community, journalism, news

CNN Shutting Down Its Facebook In Australia Shows How Removing 230 Will Silence Speech

from the liability-is-too-great dept

It remains perplexing to me that so many people — especially among the Trumpist world — seem to believe that removing Section 230 will somehow make websites more likely to host their incendiary speech. We’ve explained before why the opposite is true — adding more liability for user speech means a lot fewer sites will allow user speech. But now we have a real world example to show this.

Last month, in a truly bizarre ruling, the Australian High Court said that news publishers should be liable for comments on social media on their own posts to those social media platforms. In other words, if a news organization published a story about, say, a politician, and then linked to that story on Facebook, if a random user defamed the politician in the comments on Facebook… then the original publisher could face liability for those comments.

It didn’t take long for Rupert Murdoch (who has been pushing to end Section 230 in the US) to start screaming about how he and other media publishers now need special intermediary protections in Australia. And he’s not wrong (even if he is hypocritical). But, even more interesting is that CNN has announced that it will no longer publish news to Facebook in Australia in response to this law:

CNN says it will no longer publish content to Facebook in Australia. The decision comes after the country’s highest court ruled that media companies are liable for comments people post under articles on the platform.

Of course, this is also amusing, because CNN is owned by Turner Broadcasting, which is owned by WarnerMedia which (for the time being) is owned by AT&T… which has also been a recent critic of 230 (despite being protected by 230). Of course, in the process CNN made sure to try to blame Facebook for this decision:

After that ruling, CNN approached Facebook and asked if the tech firm would “support CNN and other publishers by disabling the comment functionality on their platform in Australia,” a CNN spokesperson said in a statement, adding that Facebook “chose not to do so.”

“We are disappointed that Facebook, once again, has failed to ensure its platform is a place for credible journalism and productive dialogue around current events among its users,” the CNN spokesperson said, adding that the media outlet will continue to publish on its own platforms in Australia.

Of course, this now raises a separate question. Remember, earlier this year, Australia passed its ridiculous (Murdoch designed) link tax, forcing Facebook to pay news organizations if they post news to Facebook.

But if, under this recent court ruling, it forces news orgs to stop posting their news to Facebook… does that mean Facebook no longer needs to pay them?

Either way, CNN’s decision to withdraw its content from Facebook in Australia is a perfect example of how increasing intermediary liability means less speech overall. Removing Section 230 wouldn’t lead to more speech online, it would lead to a lot less.

Filed Under: australia, comments, free speech, intermediary liability, risk, section 230
Companies: at&t, cnn, warnermedia