431: Barrett’s Moody Concurrence: Oddly Popular, Wholly Wrong

[Music] Welcome back to the Tech Policy Podcast. I'm Corvian Barthold. On an episode last year, I went on a short diatribe about Justice Amy Coney Barrett's concurrence in Moody vs. Net Choice. That's the one where she questioned whether the First Amendment has anything to say in situations where an algorithm or AI does how to put it. Too much work in making editorial decisions for a editor. What if a platform's algorithm, she asked, just presents automatically to each user, whatever the algorithm thinks the user will like. E.G., contents similar to posts with which the user previously engaged. And what about AI, she went on? What if a platform's owners hand the reins to an AI tool and ask it simply to remove, quote, hateful content. She worried that in these situations, perhaps, a human being with First Amendment rights has not made an inherently expressive choice. The way services, quote, "use this sort of technology," she concluded, "might have constitutional significance." As I said in my critique, there seems to be an assumption here that algorithms just spring up. You say, "Algorithm, go." It's true, I noted, that AI researchers don't know precisely how LLM's work, did social media and algorithms, and that they can't predict precisely what they will do, but they are constantly tweaking these things. Never mind, I went on, the fact that having algorithms give the user what they want is itself an expressive choice. I had a problem furthermore with the notion that a human absolutely must be in the loop for the First Amendment to have any bite. Words have semantic sovereignty. I said, quoting myself, "You have a reaction to words that you see regardless of who the speaker was or what their intention was like. The death of the author is a real postmodernist insight that, regardless of what the author intended to convey to you, you have some kind of reaction to words." There's a strong First Amendment tradition of listeners having rights to receive information. Self-indulgent self-quote over. So, I had a problem with this opinion. The opinion came out several years ago. That podcast episode was last year. What you might wonder, are we doing here now? Well, Barrett's concurrence has had a remarkable influence on wider jurisprudence in the courts, and I have so much ranting left to do. So now is as good a time as any to lay down a thorough deconstruction of the fallacies in Barrett's thinking in this concurrence. I think this kind of analysis needs to exist somewhere, so it might as well be here and now. Moody versus net choice for those of you who need a refresher. That was the Supreme Court's big social media case from the 2023 term. The decision issued on July 1st, 2024 to be precise. The case involved a pair of challenges to Texas and Florida laws, restricting how social media platforms moderate content. In the majority opinion, Justice Kagan supplied a full-throated defense of editorial discretion. When platforms compile or curate or arrange third-party content into a newsfeed, they exercise editorial judgment protected by the First Amendment. Justice Barrett joined the majority in full, but she also offered a separate concurrence that has taken on a life of its own. The majority vacated the judgments because it found that the courts of appeals had not conducted a proper analysis of the facial First Amendment challenges that had been brought. In other words, the plaintiffs had not shown that a substantial number of the laws' applications were unconstitutional, judged in relation to the statute's plainly legitimate suite. Justice Barrett elaborated on how hard she thought it would be for the plaintiffs to meet that bar and along the way she did three things. First, she picked that the role of algorithms. Justice Barrett had no problem with a social media company using an algorithm to carry out a discrete, easily understandable, expressive choice, take down all content that denies the holocaust or whatever. She thinks the First Amendment clearly applies to that. But she questioned whether the First Amendment continues to apply if the company just sets up the algorithm to, as I mentioned earlier, present automatically to each user whatever the algorithm thinks the user would like. She says that might be different. Second, AI. What if a company just has AI make most of the choices? Again, as I mentioned earlier, instructing an AI, remove hateful content, go. Barrett asks, if the AI is just relying on the data in a large language model to determine what qualifies as hateful, has a human being made enough of a choice there for the First Amendment to continue to apply. Does AI, in other words, put a gap between the human decision maker and the outcome that is big enough that the First Amendment goes away? Third, foreign ownership. Barrett suggested that if a social media platform has foreign ownership, that might mean that laws overriding the platform's editorial decisions don't trigger the First Amendment. That was a clear nod to TikTok. Her comments clearly peraged the Supreme Court's TikTok decision. This is yet another way in which Barrett's concurrence was hugely influential, but this one is not a topic for further discussion in this episode. Barrett framed all of her analysis as an exercise in just asking questions, which is interesting, giving the influence that the concurrence has had. I'm going to get to that in just a moment. Presumably to secure her vote, the majority incorporated a footnote five of its opinion, echoing Barrett's concerns, says, quote, "We do not deal here with feeds whose algorithms respond solely to how users act online, giving them the content they appear to want, without any regard to independent content standards." That footnote then goes on to preview, the fatal flaw in Barrett's analysis. It says, "Like them or low them, the community standards and community guidelines make a wealth of user agnostic judgments about what kinds of speech, including what viewpoints are not worthy of promotion." Those judgments show up in Facebook's and YouTube's main feeds. On that podcast, I mentioned at the outset, I brought up the great Supreme Court litigator Lisa Blatt and her wonderful comment that she, "All star Supreme Court practitioner does not read these concurrences. She only focuses on the majority opinions. It would appear that she is virtually alone in that practice, however, because Barrett's concurrence, at least, is popping up all over the place." First up, last fall, the ninth circuit issued a decision called Net Choice versus Bonta. This involved a challenge to Californians protecting our kids from social media addiction act. Law restricts Miner's access to algorithmically personalized feeds unless a parent opts in. Net Choice, one of the parties obviously in Moody versus Net Choice, challenged the law on First Amendment grounds. The district court relied heavily on Moody's footnote five and Barrett's concurrence to largely deny injunctive relief to Net Choice in that case. Quote, "When it comes to feeds that recommend Post-Based Soleil on prior user activity, there is no apparent message being conveyed." Then I circuit largely affirmative. termed, held that net-choice-lact-associational standing to bring as-applied challenges to the law because, quoting Barrett, "the first amendment analysis is fact-intensive and will surely vary from function to function and platform to platform." That's the part I mentioned earlier about Barrett bolstering the majority's conclusion that we have to get really granular in looking at each different aspect of a social media platform to decide if each one warns first amendment protection. But more to the point, the Ninth Circuit, citing footnote 5 and Barrett's concurrence, acclaimed that an algorithm that, quote, "response solely to how users act online, merely giving them the content they appear to want, probably is not expressive." So there's a court taking Barrett's concurrence and running with it. Next up, Garcia versus character AI. This was a case in the middle district of Florida. It's the one that prompted my discussion of Barrett's concurrence in that podcast episode last year. This is the case of a 14-year-old who tragically took his own life after interacting with a chatbot on the character AI platform. After move to dismiss the lawsuit on first amendment grounds, among other things, Judge Anne Conway denied the motion. She leaned explicitly on Barrett's concurrence. She likened character AI's outputs to an algorithm that just presents users what they want. And given that analogy, she declared herself not prepared to hold the character AI's output is speech. Judge Conway jumped over, did not grapple with, did not mention, a bunch of precedent affirming your and my and everyone else's right to information under the first amendment. Cases like Stanley versus Georgia, which says that you can possess obscene, even obscene material in the privacy of your own home because you have a right to information. Decisions like Lamont versus Postmaster General, which hold that you can receive Communist propaganda, direct from the CCP, even though the CCP has no first amendment rights, you have a first amendment right to information. I for one would humbly suggest if you have a first amendment right to receive information from a genocidal regime, you also have a first amendment right to receive information from a chatbot, but I digress. For Judge Conway, those precedents had to give way, even if only implicitly, to the sheer analytical power of Amy Coney Barrett's concurrence in Moody versus Net Choice. Finally, a trial court in North Carolina last year also relied on Barrett's concurrence. The state attorney general sued bite dance for unfair and deceptive trade practices. Bite dance raised a first amendment defense. Court denied a motion to dismiss quoting Barrett and footnote five. Court wrote that the TikTok algorithm, quote, "does not convey a message by its programmer. It simply bows to user preferences and propensities." Barrett's concurrence, I should note, is making the rounds in certain intellectual circles. There's an essay in the latest edition of National Affairs by John Errett and Brad Little John called the Post-Human First Amendment. I'm going to return to this one at the end. It shows that conservative legal thought is picking up Barrett's concurrence and running with it. The article treats the concurrence basically as grist for making the argument that we need to retreat from modern First Amendment doctrine more broadly. So Barrett's concurrence short, just asking questions, but having major effects. The rest on this notion that First Amendment protection tracks the degree of identifiable human editorial choice that sits behind an algorithm or an AI output. More human direction, more protection. Without human direction or with little human direction, the expressive power of what we're looking at doesn't matter at all. It goes by the wayside. And I now like at last getting to all of the problems here. So I'm going to come at this from three angles. They're intertwined. So you'll hear bits and pieces of each in the others. But for logical coherence, let's think of this as the legal problems, the technical problems, and the philosophical problems with what Barrett has done here. Start with the law. Barrett's analysis has no support in existing doctrine. To begin with, you can just look at the majority opinion that Barrett herself joined. The moody majority ruled that social media news feeds involve editorial judgments that are protected expressive activity. And it reached this conclusion while fully acknowledging that those feeds heavily use algorithms that prioritize content most often based on a user's expressed interests and past activities. The whole thing, the selection, the ranking, the prioritization, the presentation, it's all editorial discretion. It didn't pull out the behavioral cues and treat it as separate and apart from the editorial function. If anything, in footnote 5, the majority made pretty clear that it doesn't think such a separate product exists. It described Barrett's hypothetical and then immediately noted that no platform comes anywhere near embodying that hypothetical. The majority at least seemed to understand that every platform applies a mix of behavioral signals and content standards in tandem to create its expressive product. They are completely entangled. Moving outside the majority opinion in Moody vs. Net Choice, there is a decision that a lot of us were raising before the Supreme Court issued its decision, hurly versus Irish-American gay lesbian and bisexual group of Boston. This is the Boston St. Patrick's Day parade that seemed to let pretty much everyone march in its parade except for an LGBTQ group that it would not allow to join. Seems pretty bigoted, unsavory, not a good look. But the First Amendment doesn't require people to be virtuous. And what Hurley said in upholding the parade's First Amendment right to exclude one specific group even when it led in pretty much everyone else is that an editor doesn't need a narrow, succinctly, articulable message to be engaging in expressive activity protected by the First Amendment. To put it another way, the First Amendment does not contain a "use it or lose it trap door" as I have put elsewhere. As Hurley observed, the parade did not forfeit constitutional protection simply by combining multi-farius voices. Okay, now for the technical or technological error. Once again, Barrett's hypothetical, the feed that responds solely to how a user acts online what they seem to want or an AI that makes all the decisions without any input from a human. This is not a description of anything that exists. There's no such thing as a pure preference mirror or a untouched adenic AI. Barrett has offered us a thought experiment that is masquerading as a almost utopian technological vision. What the user wants is not a determinant concept. And at least for the time being, there is no such thing as AI's with completely independent agency. To flesh this out, let's focus on social media for the moment. A social media service has an infinite number, virtually infinite, number of things that it can serve you. It has to decide among them. And this brings us straight to the question of what is engagement because generally speaking, if it's trying to give you things that it thinks you want, that is going to be pegged to what material other people want, what material is popular. And how is that determined? Lot of choices have to be made. To be made here is popularity measured by watch time, clicks, shares, comments. How you do this is going to profoundly affect the feed. A feed optimized for watch time might tend to long absorbing or hypnotic content. One optimized for shares might tend to outrage and moorlanding nation. You might be optimizing for return visits, in which case you're going to be looking to maximize satisfaction rather than exhaustion. These aren't minor variations. They're going to produce radically different information environments. And behind it all, are choices, human choices, editorial choices made by the social media service. Put another way, what do you want? That is not a platonic concept. You want positive or negative content, stuff from your connections or more viral stuff, stuff that you engage with, or again, what's viral? How do you determine what becomes viral? Is it based on the initial engagement it gets from other users? Is it weighted in favor of the big accounts? Are negative comments weighted as good or bad? What to measure and how to measure it profoundly affects what it means to give a user what they quote unquote want? These choices have to be made at every level of the architecture. The platform is deciding, again, what gets initial visibility? Who the first audience is for a fresh post? How the virality dynamics work? Over what time horizons? How quickly does something need to catch on for it to go viral? Whether we're maximizing for engagement this minute or keeping people coming back? Are we waiting for outrage? Are we waiting for trust? I'm starting to repeat myself here. But the point is, there are basically infinite feeds within the category of give the user what she wants. Think of it as how there are infinite number sets, 1, 100, 200, 300, within the number set that's a bigger infinity of 1, 2, 3, 4, 5. Just because what the user wants is a subset of the infinite sets, it could still be infinite. There is no neutral what the user wants baseline. There is no algorithm that simply reflects the user's preferences without actually building and constituting them. To make this concrete, you might have a user who got interested in a flat-earth video on YouTube and then watched another flat-earth video. What does she want next? Does she want yet another flat-earth video? Does she want a debunking video? Does she want a video about a different conspiracy? Does she actually just want to learn about astronomy more broadly? Maybe actually there's something completely unrelated that you will find even more engaging? All of those are plausible. The algorithm has to choose. And YouTube actually provided a really concrete demonstration of this. As we talked about on an episode on the show way, way, way back in 2021 that I will put in the show notes, YouTube realized that it had a problem with sending people down flat-earth rabbit holes via its recommendation system. So YouTube went through a lot of effort to change its algorithm. The flat-earth videos were still on the platform. Some users were still watching them, but YouTube tried to stop actively funneling users into them. And what happened? To be clear, I cannot speak for every individual person who was seeking flat-earth content. But on the whole, YouTube did not have a problem in continuing to find other different ways to provide the user what she wanted. Daily average user time went up between 2019 and 2025. Add revenue grew over that period. And the overall number of users of the platform continued to grow. YouTube did not reduce engagement, but it did consciously redirect it. Users still got, in Barrett's phrase, whatever the algorithm thinks the user will like, but they got a different version of like. One that involved fewer rabbit holes. In short, maximizing user engagement can maximize extraordinarily different things. Another important point. These highly malleable feedback loops that I've just been discussing are not limited to individual users. A platform is a complex adaptive system. It's algorithmic architecture functions as perhaps initial conditions. These design choices shape who joins the platform, which shapes what gets posted, which shapes who else joins, which shapes the norms of the platform, which shapes its politics, and small differences in how an algorithm waits engagement across all of this, produces over time radically different communities, different norms, different politics. There is an entire ecology of discourse, if you will. Who is on the platform, what they want to talk about, how they talk about it, who eventually leaves? Again, with tons of choices by the platform sitting behind it. Think of it this way. The question what the user wants has behind it, the incredibly important subsidiary question, which user? So for example, in June 2020, TikTok users coordinated a campaign to reserve free tickets to a Trump rally in Tulsa with no intention of attending. The result, I mean, the causal connection here is debated, but a half empty arena and political embarrassment for the administration. Some of the TikTok users saw this as brilliant digital activism, free speech at its most creative, and others saw deceptive manipulation and perhaps even an undermining of the political process. So in letting the algorithm serve up content, in this case, hey, ha, ha, look at the way that we are messing up this political rally. The question is, whose preferences do you serve? You're giving some people what they want and other people not what they want. Are you serving the activists? Are you serving the rally goers and Trump supporters? After this happens, who gets offended and stops using your product? Who thinks the product is awesome and stays? Suddenly, an algorithm that, quote, gives the user what she singular wants doesn't seem like a straightforward proposition. Similarly, consider Blue Sky. As I always say on the show, I like Blue Sky. I hope it succeeds. But it tends to have a problem with users who post aggressive, scolding replies. A lot of accusation and condemnation, pylons. This is a thing. And Normie lives on the platform, note sometimes, that this creates an exhausting environment sometimes. And then Cruz, prominent user, announced once that he was taking a break because the platform was making him sad and angry. Sarah Jung observed that these dynamics was causing a lot of power users to leave what with all the dog piling. So again, we have a choice that the platform needs to make. It can intervene and tamp down on the scolds and try to promote thoughtful posters and de-amplify hostility. Or it can let the dynamic play out in the name of giving users who are posting the scolding replies and what they want. But what it cannot do is simply give the user what she singular wants. There is no such thing. It has to serve some users' preferences at the expense of others with knock on consequences for the larger environment and fundamentally what the platform is. So Barrett gives us this simple binary with human editorial choices that are inherently expressive on the one hand and feeds that merely mirror human preference. differences on the other. But it's a false dichotomy. There is no such thing as the second one. Now from my third prong, the philosophical fallacy here, even if social media services could be as automated as Justice Barrett suggests they could be, her concurrence would remain wrong at a deeper level, at a philosophical level. And here we can shift in perhaps fruitfully to her AI point. She says, "What of a platform's owners hand the reins to an AI tool and ask it simply to remove hateful content?" Again, implies that this might not be first amendment protected decision. But notice what the humans in this hypothetical did. They made a choice to remove hateful content. Not misleading content, not sexually explicit content, or not a choice to not remove hateful content. Here they made a very specific choice to remove a very specific category. And obviously, within that category, there are a bunch of judgment calls about what counts as hateful is criticizing religion, hateful is harsh political rhetoric, hateful, etc., etc. And Barrett's point seems to be that it's a problem if a human is not making granular decisions at that level. But the fact has not changed that the system as a whole, whether by using an LLM or not, is expressing a viewpoint about what speech the platform values. The top-line choice there still matters. In my mind, it's really not different from the owner of a newspaper hiring an editor and telling that editor, "Publish rigorous newsworthy pieces." Clearly, the publisher has given the editor vast discretion. But I'd like to think that nobody would assume that the newspaper had thereby abandoned its First Amendment rights. The publisher chose an editor. The publisher chose the instruction. The publisher is going to be held responsible in the public eye for what the newspaper does. The publisher could fire the editor if the newspaper got enough heat. So what we see is that the delegation is not some kind of abdication in a sense that should have First Amendment balance. What seems to be going on here is that Justice Barrett has a double standard. Even with the editor in that hypothetical, there's a human in the loop. And having a human in the loop seems to be what really matters. But actually, I would put it to you that even that is a false dichotomy. If we were to discover that the New York Times was selecting op-eds by coin flip, I don't think that that would give a state regulator standing to override the New York Times's First Amendment right to print whatever the hell op-eds it feels like printing. Yes, this would be an eccentric editorial policy, but an editorial policy it would remain. The First Amendment does not inquire into the quality of the editorial judgment. A related point is that social media platforms seem to be faltered, not just by Justice Barrett, but more broadly for wanting to maximize engagement, as if newspapers and television networks and book publishers are not all very much making decisions driven in large part by what their audiences want and what will keep their audiences coming back. These people are tracking their readership and their ratings and their sales. I assure you, in this context, it's much easier to see that being responsive to an audience does not strip you of First Amendment protection. If the newspaper publisher hires the editor and tells her "run whatever sells papers," again, a huge delegation of discretion, something very much driven by audience behavior, the publisher clearly is not going to be able to tell you what is going to be in her own paper from day to day, but she has not thereby forfeited her editorial discretion or her First Amendment protection. The analogy, in my mind at least, holds quite strongly for the social media platform that waits what it does toward showing users material that will maximize engagement. Barrett's concurrence, not great, has had a lot of influence. I wish people wouldn't take it so seriously. With that, let me take a slight detour. Another topic I like to ramble on about. I've done it in various places. I did it just recently with the bright young kids at Berkeley Laws, Federalist Society chapter. Lives just don't seem to know what they want to do with technology. This is a broad and interesting topic that has many more facets than what I'm going to briefly touch on here. But for present purposes, let's just focus on the fact that the moodiestision is the product of an already stale, maybe even obsolete movement on the right to open up platforms and make them anything goes. The Texas and Florida laws that gave rise to movie versus net choice stood on this animating belief of anti-conservative bias among the major platforms. The remedy that Texas and Florida wanted was greater state control over the feeds, which in turn required narrowing social media companies' first amendment rights. Texas Barrett's concurrence has the feel of a compromise in service of Texas and Florida's ends. She didn't go full Alito concurrence in Moody and say, "We can just treat the platforms essentially as common carriers and force them to carry all the speech that Texas and Florida is demanding." But she did create a potential narrow passage the states could get through where they could find situations where they could coerce platforms. And since then, conservatives have moved on to other concerns like child safety online. In part because they actually ended up getting in the market. They got what they wanted good and hard. Now Barrett's concurrence is sitting there able to do mischief in ways unrelated to the initial concern that got Texas and Florida moving in the first place. So let me explain. As we all know, Elon Musk buys Twitter and turns it into X and strips back the content moderation and changes the algorithm. And at the time, conservative voices celebrate this and see it as the end of a quote-unquote big tech censorship. And since then, those of us who, fair to say, had a more sophisticated understanding of trust and safety and content moderation, we've watched as certain conservatives have learned belatedly what content moderation actually does. So Christopher Rufo, the right-wing activist, he is one guy who has spoken with a lot of candor here. He has complained recently about how X's algorithm has hijacked the rights public discourse. He says that Musk's governance of X has "detached reach from quality, incentivizing narratives that earn clicks but fail to resemble a persuasive or even meaningful public argument." He is talking about conspiracy theories, about Charlie Kirk's assassination involving Israeli intelligence. He's talking about Candace Owens. He is talking about anti-Semitic content, Tucker Carlson, all of the most unseemly fever swamps of far-right discourse. And he says that the result is an epistemic fork. The conservative mind, he says, has been co-opted by the ex-algorithm and productive discussion on the right, he says, has become meaningless and it's made action on the right impossible. He is not alone, by the way. Vivek Ramaswamy announced that he is logging off ex because of its quote, warped projection of reality. Raheem Kasam. Philly with Breitbart has called ex a quote post apocalyptic cesspool of bots and political illiteracy. It's making everyone dumbery says there are other examples. Now, let me put it to you. What is the complaint here? The complaint, of course, is that the ex-algorithm is doing the wrong kind of giving the user what she wants. It is an acknowledgement that there are many kinds of giving the user what she wants and that this one is a bad one. Rufo writes, every algorithm has a master, in this case, Musken is team. Given that their vision for ex has not been realized, they would be wise to revise the algorithm. Revise the algorithm. It is a call for editorial judgment. It is an understanding that the algorithm is the editorial policy. It is a belated realization that justice, Barrett was wrong. I know she was just asking questions, but wrong to even act like these were the simple questions to be asking. So that's interesting to me. And site note. Having noticed that Musk is doing a very crass, simplistic, toxic, brute form of giving the user what she wants. Query, whether these people or other figures on the political right, would be happy with judges saying, ah, ah, ex is now, it is our example of the pure algorithm giving the people what they want, just as justice, Barrett, prophesied, we would find one in the wild eventually. Blue states can come in and regulate X. Don't think that's what they'd want. All right, so sticking with my theme of conservatives, not really knowing what they want, just in closing, I want to return, as I said, I would to John, Erich and Brad, little Johns essay in National Affairs, the post human first amendment. I actually heard a version of the argument in this essay on a panel I did with Joel Thayer that I think will be our next episode to drop. So I'm seeing this essay gained traction. The basic idea is that the founders understood speech as a natural right grounded in human nature, that the 20th century, first amendment jurisprudence took a bunch of wrong turns with corporate speech rights and expressive conduct in particular. And that it would be absurd to give first amendment rights to AI and that therefore courts need to retrench certain aspects of this are pretty wild to me in particular, the corporate angle. The authors say that the Supreme Court should reconsider citizens united, which is very interesting, citizens united, being historically a bed and war of the left. I'm not going to get into a detail, take down of this essay or whatever, because it's not my purpose here, but I will point out the essay seems to mistake citizens united as the decision that created corporate first amendment rights. It did no such thing. Citizens United is about political spending. There are a lot of decisions like a lotty or New York Times versus Sullivan or Miami Herald versus Torneo, which either explicitly hold or have as a fundamental premise that corporations have first amendment rights to express themselves. What I will note though, is that this attack on corporate speech rights is very conservatives don't know what they want, ties into my theme completely. Do you really want the government coming in and pushing around Fox News or the Wall Street Journal op-ed page or I guess the Federalist Society? That seems like a bad idea. And the answer seems to be, well, that might be worth it. If we could get at the social media companies and the AI companies and make sure that there is not a first amendment right in LLM outputs or in social media feeds, to which I then have the same question, have you forgotten that social media was a powerful tool for conservative voices to go around the legacy media? And are you really blind to the huge potential of AI to serve conservative cultural ends? If you're disillusioned with the public school system, the AI that can teach your kid about Dante and Thomas Aquinas could be a huge boon. If you strip first amendment rights from all of this in a misguided attempt to take us back to tradition or whatever, you are inviting the state to dictate what outputs are allowed on the AI. Conservatives, whether they understand this or not, are still very much a counter cultural force. They actually have good reason to be champions of the right of anonymous speech and the right of the citizen to receive information and to have forbidden thoughts. Maybe I just say that as somebody who's living in California, maybe there's a false sense of security if you live in Tennessee or something. I don't know. It all seems just very, very misguided to me. The Trump in the essay seems to be whatever it's worth it. If we can go back to originalism and there again, I'm not clear on how that's supposed to work. Yes, it's true. We didn't even get around incorporating the first amendment until the 20th century. Much of first amendment precedent was built up from the 1950s onward. First of all, you're making an almost Jacobin case to discard starry decisis and respect for precedent and any sense of Burkeen conservatism and just tear down broad swaths of first amendment precedent and take us back to year one. You sound like Robespierre. Second, as I've discussed on the show before with my podcast BFF Arikone, there's a strong case to be made that the first amendment has simply kept up with technology in the same way that say the fourth amendment or the second amendment has. We have adopted rules that make sense in an era of mass media, universal suffrage, advanced information technology. This originalism push is a bit like saying that the second amendment needs to be limited to muskets. I actually think that many of the founders, if they could be pulled on the subject, not that this would be reliable for much of anything, but those guys liked them some political dissent and some fractiousness and they would look and approve of the ability for shit posters to malign the government on social media or to hone their potentially seditious arguments with an AI chatbot. Most of them were not cool with the alien and sedition acts mind you. Generally speaking, those were a stain on our history. Nor did they treat speech as some kind of high-minded exercise in promoting human flourishing or some nonsense like that. Free speech in the founding era meant ruthlessly taking the piss out of your political opponents in anonymous pamphlets and outraging each other and getting into tabloid level fights. Anyway, I'm sorry, I suppose that was all a bit scattered, but the upshot is that there is not some idyllic speech environment of the past to return to and tearing down social media or AI wouldn't get us there if there was. Justice Barrett's concurrence is misguided for all kinds of reasons, but it wouldn't be a promising blueprint for reform even if it wasn't. Thank you so much if you have followed me on this journey. This may seem like a narrow time. topic. But as I hope I suggested sufficiently at the end, it has brought implications. We're going to be talking about the backlash to technology and all the bad ideas that surround it plenty on the show we have in the past. We will in the future. This has been one segment in that ongoing debate. I love a lot more to say on it, but I have said enough today. This has been Tech Policy Podcast. I am your host, Corp. Barthold. If you enjoy conversations like this one, please do go give the show a five-star rating for where every you listen. And while you go do that, I will get started on the next one. Thank you all. Until next time. The Tech Policy Podcast is produced and distributed by Tech Freedom, a nonpartisan, nonprofit think tank in Washington, DC. To learn more about our work, make a tax deductible donation or find other episodes, find us online at techfreedom.org. [Music]