From Gods to Google
abstract. The First Amendment has become a significant barrier to sensible technology regulation. The conventional explanation for this is the Court’s deregulatory turn in free-speech law. But the Lochnerization story is incomplete. The Court’s profound solicitude for religious speakers plays a central role in the current digital-free-expression landscape. By protecting the speech of certain religious dissidents, the Court has created a set of constitutional entitlements that logically extend to technology firms. Along the way, the Court has eroded its ability to apply the First Amendment sensibly to novel technologies.
This Feature draws the doctrinal through line from gods to Google. We first sketch the basic contours of today’s technology regulation and explain why it is vulnerable to First Amendment challenge. We then give an overview of free-speech case law that develops what we call the oppressed-speaker paradigm. The Roberts Court has been motivated not just by free-market zeal but also by the trope of a persecuted religious minority standing fast in the face of a domineering and majoritarian regulator.
We pay special attention to 303 Creative LLC v. Elenis, identifying several doctrinal defects likely to have an impact on technology regulation. Technology firms have wasted no time in relying on 303 Creative to challenge a variety of new laws—most notably in last Term’s Moody v. NetChoice, LLC, which involved free-speech challenges to the regulation of internet platforms. The Court in Moody, however, sidestepped or ignored the most serious implications of 303 Creative.
It may be tempting to think the Court could reconcile the two cases by distinguishing religious speakers from platforms. But doing so would impermissibly enshrine viewpoint and speaker discrimination into free-speech law. The Court thus confronts a conundrum of its own making: either (1) apply the principles developed for religious speakers to new technologies and expose a wide range of technology policy tools to constitutional attack, or (2) create special rules for religious speakers, which would violate the Court’s own notions about viewpoint and speaker neutrality. A principled resolution of this conflict cannot be that free-speech law affords special protection for religious speakers. The Roberts Court must find legitimate and coherent limiting principles for the First Amendment landmines it has laid.
authors. Rebecca Aviel is Professor of Law and Maxine Kurtz Faculty Research Scholar, University of Denver Sturm College of Law. Margot E. Kaminski is Professor of Law, University of Colorado Law School; Faculty Director, Privacy Initiative at Silicon Flatirons Center. Toni M. Massaro is Regents Professor of Law Emerita and Dean and Milton O. Riepe Chair in Constitutional Law Emerita, University of Arizona James E. Rogers College of Law. Andrew Keane Woods is Milton O. Riepe Professor of Law and Distinguished Legal Scholar, University of Arizona James E. Rogers College of Law. We thank the Yale Law Journal editors for their care and rigor throughout the editorial process.
Introduction
Digital technologies raise a host of public-policy puzzles. The use of artificial intelligence (AI) in both the private and public sectors raises concerns about the fairness, transparency, and accountability of these systems.1 As social-media platforms have become our “modern public square,”2 critics have decried platforms’ content-moderation practices as censorship.3 The explosion of “surveillance capitalism” imperils citizens’ privacy.4 Law-enforcement officers seek access to encrypted devices and services,5 while the U.S. government perceives a national-security risk in the possibility that foreign-owned technology companies will export users’ sensitive personal data.6 And then there are the children.7 In 2023, the U.S. Surgeon General issued an advisory detailing “ample indicators” that social media poses “a profound risk of harm to the mental health and well-being of children and adolescents.”8 As one doctor put it, “The internet is a giant hypodermic, and the content, including social media like Meta, are the psychoactive drugs.”9
Across these different domains, there is both bipartisan agreement that regulation should be enacted and a rough consensus about the relevant set of policy tools available to do so. The problem is that these tools are vulnerable to powerful First Amendment challenges. Because digital technologies involve “images, words, symbols, and other modes of expression,”10 nearly every regulation of digital technologies could be characterized under current law as a restraint on speech.11 This is partly a result of what scholars call the Lochnerization of the First Amendment, driven by the Supreme Court’s free-market skepticism of legislative policymaking.12 But Lochnerization is only part of the picture.
In this Feature, we show that the Court’s profound solicitude for religious speakers is a central but underappreciated part of the current digital-free-expression landscape. By protecting the speech of certain religious dissidents, the Court has created a set of constitutional entitlements that logically extend from gods to Google. Along the way, the Court has eroded its ability to apply the First Amendment sensibly to novel technologies. The Court cannot avoid this problem by simply distinguishing religious speakers from technology speakers. Doing so would violate a core free-speech principle prohibiting viewpoint discrimination; indeed, this Court has admonished the political branches that they are “prohibited” from enacting “restrictions distinguishing among different speakers, allowing speech by some but not others.”13 The Court thus confronts a conundrum of its own making: either (1) apply the principles developed for religious speakers to new technologies, rendering a wide range of technology policy tools vulnerable to constitutional attack, or (2) create special rules for religious speakers, which would violate the Court’s own notions about viewpoint and speaker neutrality.
Two recent Supreme Court cases illustrate the dilemma. In 303 Creative LLC v. Elenis, the Court excused a Christian website designer from compliance with a state public-accommodations law. The Court held that requiring the website designer to create a wedding website for a same-sex couple compelled “pure speech” in service of government ideology14—as if the government were forcing schoolchildren to pledge allegiance to the flag, like in West Virginia Board of Education v. Barnette.15 The Court did not consider whether the state could demonstrate compelling reasons for its nondiscrimination law. It simply announced a categorical prohibition against government compulsion of “pure speech.”16 And because 303 Creative was decided on free-speech rather than free-exercise grounds, its reach logically cannot be limited to religious speakers.17
Recognizing this, technology companies immediately seized upon the case to challenge state regulation of internet platforms in Moody v. NetChoice, LLC.18 Moody involved two state laws that restrict platforms’ ability to control how third-party content is presented to other users—including decisions to filter, prioritize, label, and remove posts—and require platforms to provide individualized explanations of content-moderation decisions.19 Paul Clement, representing the platforms, argued that these state laws sought to compel speech in exactly the way that 303 Creative flatly foreclosed. Clement’s opening statement revealed how central 303 Creative was to the platforms’ argument: “[Y]ou cannot have the forced dissemination of third-party speech . . . . And Reno [v. ACLU] and 303 Creative make clear those principles are fully applicable on the Internet.”20 If the Court stands behind 303 Creative and holds fast to its views about speaker neutrality, then the protections it has developed for religious speakers must extend to technology companies.
The Becket Fund for Religious Liberty, in an amicus brief supporting neither party, offered the Court a starkly different path.21 It urged the Court to cabin its religious-speaker cases by explicitly granting special protection to religious speakers only.22 But carving out special rules for religious speakers would inject a new and pernicious dynamic into free-speech doctrine, casting doubt on decades of case law emphasizing the importance of speaker and viewpoint neutrality.23 It is the Free Exercise Clause, with its textual commitment to the protection of religion, that offers special consideration to faith-based reasons for acting or refusing to act.24 In the speech realm, the government must be neutral toward a range of ideologies and viewpoints, including attitudes about religion.25 It can neither disfavor nor favor religious speech. The First Amendment’s nearly absolute prohibition on viewpoint discrimination means that putting religious speech on a higher plane than nonreligious speech (or antireligious speech) is no more acceptable than favoring liberal speech over conservative speech.26
Given the express arguments made to the Court in Moody, the Justices could not have missed the question: are technology companies protected by the absolutist principles embraced in 303 Creative, or is that approach reserved for speakers with religious convictions? Instead of answering the question, the Moody Court blinked. All nine Justices agreed to vacate and remand the cases, holding that the courts of appeals below had failed to apply the correct standard for facial challenges under the First Amendment.27 They flatly ignored the technology companies’ arguments that 303 Creative’s firm assertion that website design is “pure speech” was either controlling in the platform cases or somehow distinguishable on other grounds, including on the ground that the website designer in 303 Creative was a religious speaker.28 Litigants and lower courts are now left with the difficult task of squaring Moody with 303 Creative and earlier religious-speaker cases.29 All remain viable cases in the Court’s new doctrinal thicket and imperil sensible technology regulation as much as the case law driven by Lochnerian impulses.
The significant implications of the religious-speaker case law for modern technology law have not been previously identified. This likely is because the topics of religious speakers and technology regulation have long been viewed as distinct rather than potentially related phenomena. Although technology-law scholars and policymakers are acutely aware of the deregulatory consequences of recent First Amendment cases in general,30and some commentators even anticipate application of First Amendment protections to nonhuman technology speakers,31 none has examined the significant role that religious-speaker protection per se has played in expanding this toolkit. Meanwhile, First Amendment scholars have criticized the Court’s religious-speaker cases, but largely on their own terms.32 No scholar has thoroughly explored how the cases also may jeopardize regulatory responses to the harms of digital technology. We bridge these gaps here.
This Feature proceeds in four Parts. Part I sketches, in broad strokes, what current technology regulation looks like and why. This includes platform regulation, data privacy laws, national-security laws, regulations to protect children, and laws governing AI. This summary sets the stage for our discussion of why the Court’s religious-speaker cases are particularly dangerous for much of contemporary technology law.
Part II provides an overview of the First Amendment cases leading up to 303 Creative that illustrate what we call the oppressed-speaker paradigm. The Roberts Court is motivated by the trope of a persecuted religious minority standing fast in the face of a domineering and ideologically majoritarian regulator. It often invokes Barnette, which held that Jehovah’s Witness schoolchildren have a free-speech right not to recite the Pledge of Allegiance or salute the flag upon pain of expulsion from public school (and criminal sanctions for their parents).33 Yet the compelling Barnette narrative of conscience and courage often is ill-suited for other contexts and speakers, including ones that involve complex regulation of emerging technologies. After all, these speakers generally are not individuals, not conscience-motivated, not relatively powerless, and not subject to comparable majoritarian pressures expressed through coercive government power. Indeed, in some cases, the government regulator may be acting to defend other, minoritized speakers from censorship and undue coercion by the private technology speakers.
Part III turns to 303 Creative. The majority opinion makes five problematic moves: (1) applying free-speech principles rather than free-exercise principles to Colorado’s public-accommodations law; (2) treating the design of wedding websites as “pure speech” rather than a combination of conduct and speech; (3) characterizing the website as the speech of the vendor rather than the speech of the customer; (4) assuming the antidiscrimination mandate was unduly coercive without actually assessing the magnitude of the compliance burden; and (5) concluding that the antidiscrimination mandate was categorically unconstitutional without examining the state’s justifications. For each doctrinal move, we explain the potential threat to sensible technology policy.
Finally, in Part IV, we parse Moody and its implications. On the one hand, the Court’s application of the First Amendment to technology platforms was subtler and more sophisticated than its treatment of website design in 303 Creative. This is a welcome development. But Moody’s refusal to acknowledge, much less explain, how the two cases interact creates a doctrinal mess for lower courts and regulators. Squaring the more nuanced approach on display in Moody with the “pure speech” absolutism of 303 Creative will not be an easy task.
We argue that the Court must confront both the doctrinal excesses and the internal conundrum within its recent cases. We maintain that free-speech law normatively and analytically should not grant special protection to religious speakers. The proper doctrinal through line must be from gods to Google, and back. If the Court is unwilling to embrace the “pure speech” construct when deployed by technology companies, it must reevaluate its overprotective treatment of religious speakers. Or the Court must clearly provide guidelines to determine how the doctrine might take into account context-specific distinctions that do not depend implicitly on a speaker’s religious views. It must offer, in other words, First Amendment limiting principles derived from the role that speech ought to play in our democracy. This will require far more care and nuance than evidenced by current First Amendment doctrine.
See Ginia Bellafante, If Your Child Is Addicted to TikTok, This May Be the Cure, N.Y. Times (Nov. 17, 2023), https://www.nytimes.com/2023/11/17/nyregion/tiktok-social-media-children-addiction.html [https://perma.cc/4MCK-XYTU] (noting claims that the content and addictive properties of social media are harming children). For a discussion of social media and addiction, see generally Matthew B. Lawrence, Addiction and Liberty, 108 Cornell L. Rev. 259 (2023); and Matthew B. Lawrence, Public Health Law’s Digital Frontier: Addictive Design, Section 230, and the Freedom of Speech, 4 J. Free Speech L. 299 (2024) [hereinafter Lawrence, Public Health].
Social Media and Youth Mental Health: The U.S. Surgeon General’s Advisory, U.S. Dep’t of Health & Hum. Servs. 4 (2023), https://www.hhs.gov/sites/default/files/sg-youth-mental-health-social-media-advisory.pdf [https://perma.cc/6BDW-V7ZV]; see also Albertina Antognini & Andrew Keane Woods, Shallow Fakes, 128 Pa. St. L. Rev. 69, 94-116 (2023) (identifying a taxonomy of harms that stem from social media, including the harms experienced by teenage users in particular).
Matt Richtel, Is Social Media Addictive? Here’s What the Science Says., N.Y. Times (Oct. 25, 2023), https://www.nytimes.com/2023/10/25/health/social-media-addiction.html [https://perma.cc/9EFL-2XCU].
For a very recent example, look at the First Amendment challenges raised by technology companies against the California Age-Appropriate Design Code Act (CAADCA), a law enacted “with the express aims of promoting robust online privacy protections for children under the age of eighteen.” NetChoice, LLC v. Bonta, 113 F.4th 1101, 1108 (9th Cir. 2024).
Seee.g., Amanda Shanor, The New Lochner, 2016 Wis. L. Rev. 133, 134-35 (discussing the development and consequences of “a growing constitutional conflict between the First Amendment and the modern administrative state”); Robert Post & Amanda Shanor, Adam Smith’s First Amendment, 128 Harv. L. Rev. F. 165, 182 (2015); cf. Jeremy K. Kessler, The Early Years of First Amendment Lochnerism, 116 Colum. L. Rev. 1915, 1917 (2016) (noting that, although claims are being made that the First Amendment is being used to resurrect “economically libertarian substantive due process jurisprudence of the early twentieth century,” such concerns can be traced back to the 1930s and 1940s); Genevieve Lakier, The First Amendment’s Real Lochner Problem, 87 U. Chi. L. Rev. 1241, 1246 (2020) (arguing that the Lochner-like nature of contemporary free-speech law stems from an interpretation of the right as a “strong but limited negative autonomy right” that “guarantees freedom from intentional government interference with an individual’s autonomy, but . . . provides almost no protection whatsoever from private interference and constraint”).
603 U.S. 707 (2024). Both the petitioners and respondents discussed 303 Creative in their briefs. Brief for Petitioner at 38, Moody, 603 U.S. 707 (No. 22-277); Brief for Respondents at 47-48, Moody, 603 U.S. 707 (No. 22-277). Twenty-one amicus briefs cited or discussed 303 Creative. See, e.g., Brief for Electronic Frontier Foundation et al. as Amici Curiae in Support of Petitioners (No. 22-555) and Respondents (No. 22-277) at 21, Moody, 603 U.S. 707 (Nos. 22-277, 22-555); Brief for American Principles Project as Amici Curiae Supporting Petitioners in No. 22-277 and Respondent in No. 22-555 at 13-16, Moody, 603 U.S. 707 (2024) (Nos. 22-277, 22-555). It was also front and center in oral arguments. Transcript of Oral Argument at 63, 102, Moody, 603 U.S. 707 (No. 22-277).
The Becket Fund for Religious Liberty noted that “NetChoice and CCIA rely heavily on cases involving religious speakers, including 303 Creative, Church of Lukumi Babalu Aye, Inc. v. City of Hialeah, 508 U.S. 520 (1993), Masterpiece Cakeshop, Ltd. v. Colorado C.R. Comm’n, 138 S. Ct. 1719 (2018), McCullen v. Coakley, 573 U.S. 464 (2014), NIFLA v. Becerra, 138 S. Ct. 2361 (2018), Reed v. Town of Gilbert, 576 U.S. 155 (2015), Watchtower Bible & Tract Soc’y of N.Y. v. Village of Stratton, 536 U.S. 150 (2002), West Va. State Bd. of Educ. v. Barnette, 319 U.S. 624 (1943), and Wooley v. Maynard, 430 U.S. 705 (1977).” Brief Amicus Curiae of the Becket Fund for Religious Liberty in Support of Neither Party at 4, Moody, 603 U.S. 707 (Nos. 22-277, 22-555).
See Citizens United v. FEC, 558 U.S. 310, 340 (2010) (“Quite apart from the purpose or effect of regulating content, moreover, the Government may commit a constitutional wrong when by law it identifies certain preferred speakers.”); see also Michael Kagan, Speaker Discrimination: The Next Frontier of Free Speech, 42 Fla. St. U. L. Rev. 765, 766 (2015) (“With Citizens United, the Court for the first time gave full-throated articulation to the principle that discrimination on the basis of the identity of the speaker is offensive to the First Amendment, even when there is no content discrimination.”); Sonja R. West, Press Exceptionalism, 127 Harv. L. Rev. 2434, 2436 (2014) (“For constitutional purposes at least, it is entirely irrelevant to courts whether the speakers are members of the press or whether they are actively pursuing the news.”); Richmond Newspapers, Inc. v. Virginia, 448 U.S. 555, 573 (1980) (plurality opinion) (“[M]edia representatives enjoy the same right of access as the public . . . .”); Pennekamp v. Florida, 328 U.S. 331, 364 (1946) (Frankfurter, J., concurring) (stating that “the liberty of the press is no greater and no less” than that of everyone else (quoting R v. Gray [1900] 2 QB 36 at 40 (Eng.))).
See, e.g., Robert Post, Public Accommodations and the First Amendment: 303 Creative and “Pure Speech,” 2023 Sup. Ct. Rev. 251, 301 (explaining that “while claims of conscience may be relevant to Free Exercise jurisprudence, they have no natural home in free speech doctrine” and that “[b]y improperly transposing intuitions about religious freedom into the quite different context of freedom of speech, the Court in 303 Creative creates doctrinal chaos”); Ashutosh Bhagwat, The Conscience of the Baker: Religion and Compelled Speech, 28 Wm. & Mary Bill Rts. J. 287, 310 (2019) (arguing, in the interval between Masterpiece Cakeshop and 303 Creative, that “pure compelled-speech claims are fundamentally conscience-based” and therefore “should be rooted in the constitutional provision which protects conscience rather than the provision which protects democracy,” meaning “the Free Exercise Clause, not, as current doctrine suggests, the Free Speech Clause”).
Just as “giving offense is a viewpoint,” Matal v. Tam, 582 U.S. 218, 243 (2017), so too is giving glory to God. Speech that conveys the speaker’s sense of religious obligations or divinely revealed truth expresses a viewpoint about the world and one’s relation to it. The fact that religious speech encapsulates a potentially limitless range of different faiths, each with their own view of the “duty which we owe to our Creator, and the manner of discharging it,” Everson v. Bd. of Educ., 330 U.S. 1, 64 (1947), does not diminish the extent to which religious speech expresses a viewpoint for the purposes of free-speech analysis.
After repeatedly specifying that it seeks special status for religious speakers, the Becket brief closes by urging the Court to “distinguish between the claims of conscience and the claims of commerce as they apply to speech.” Brief Amicus Curiae of the Becket Fund for Religious Liberty in Support of Neither Party, supra note 21, at 8. We note that this is a very different proposition—a pluralist proposal to privilege all claims of conscience, whether secular or religious, over merely commercial interests. Were this in fact the position being advanced, we wouldn’t necessarily disagree—at least at this level of abstraction. There is a long tradition in free-speech jurisprudence of treating commercial speech as lower on the speech hierarchy than core political speech. See, e.g., Cent. Hudson Gas & Elec. Corp. v. Pub. Serv. Comm’n, 447 U.S. 557, 564-66 (1980) (setting forth the intermediate-scrutiny test for commercial speech). The problem is that 303 Creative itself did a great deal to unsettle this distinction. The indisputably commercial context in which it arose suggests that the Court was uninterested in preserving a more deferential approach to state regulation of the marketplace. Its failure to engage sufficiently with the fact that the plaintiff was a business open to the public is one of its most troubling defects, as we explain in Part III.
Justice Kagan noted that “NetChoice chose to litigate these cases as facial challenges, and that decision comes at a cost.” Moody, 603 U.S. at 723. She further explained that the applicable standard for facial challenges in free-speech cases is whether “a substantial number of [the law’s] applications are unconstitutional, judged in relation to the statute’s plainly legitimate sweep.” Id. (quoting Ams. for Prosperity Found. (APF) v. Bonta, 594 U.S. 595, 615 (2021)) (citing United States v. Hansen, 599 U.S. 762, 770 (2023)). She continued:
The first step in the proper facial analysis is to assess the state laws’ scope. What activities, by what actors, do the laws prohibit or otherwise regulate? . . . The next order of business is to decide which of the laws’ applications violate the First Amendment, and to measure them against the rest.
Id. at 725.
See, for example, the Ninth Circuit’s recent opinion in NetChoice, LLC v. Bonta, 113 F.4th 1101, 1108 (9th Cir. 2024), which evaluates the CAADCA and treats Moody largely as a call for the correct standard for facial challenges. The Ninth Circuit went ahead to analyze the law under the First Amendment, applying strict scrutiny to its provisions on impact assessments. Id. (“[W]e conclude that this oversight [in not applying the correct approach to facial challenges] did not cause any error in the district court’s analysis of the CAADCA’s [Data Protection Impact Assessment (DPIA)] report requirement. That is because the DPIA report requirement, in every application to a covered business, raises the same First Amendment issues.” (citation omitted)).
For an example of a recent First Amendment case that has alarmed some technology-law scholars and policymakers, see Sorrell v. IMS Health Inc., 564 U.S. 552, 558-61 (2011), which addressed a data privacy law of sorts. See infra text accompanying notes 159-165 (discussing Sorrell’s legacy in later Supreme Court cases); see also, e.g., Neil M. Richards, Why Data Privacy Law Is (Mostly) Constitutional, 56 Wm. & Mary L. Rev. 1501, 1507 (2015) (arguing that data privacy law generally does not implicate First Amendment concerns because “the question is not the ‘speechiness’ of the human activity being regulated, but the purpose and effect of the government regulation”); Ashutosh Bhagwat, Sorrell v. IMS Health: Details, Detailing, and the Death of Privacy, 36 Vt. L. Rev. 855, 879-90 (2012) (“[T]he First Amendment provides full constitutional protection to disclosures of even personal data . . . .”); Gautam Hans, No Exit: Ten Years of ‘Privacy vs. Speech’ Post-Sorrell, 65 Wash. U. J.L. & Pol’y 19, 23 (2021) (contending that privacy laws “should not automatically fall into the category of content-based regulations” but could nonetheless satisfy “strict scrutiny”); Kyle Langvardt, Crypto’s First Amendment Hustle, 26 Yale J.L. & Tech. 130, 132-33 (2024) (acknowledging and rejecting the argument that cryptocurrencies should receive First Amendment protection); Alan Z. Rozenshtein, Silicon Valley’s Speech: Technology Giants and the Deregulatory First Amendment, 1 J. Free Speech L. 337, 357 (2021) (acknowledging that technology companies may use the First Amendment for deregulatory purposes and urging doctrinal distinctions between companies arguing for users’ First Amendment rights and those arguing for their own rights); James Grimmelmann, Speech Engines, 98 Minn. L. Rev. 868, 874 (2014) (proposing that a search engine is not a “conduit” or “editor” but rather an “advisor”); Jeff Kosseff, First Amendment Protection for Online Platforms, 35 Comput. L. & Sec. Rev. 199, 200 (2019) (finding that the First Amendment only offers platforms “limited protection[s]” when compared to Section 230); Tim Wu, Machine Speech, 161 U. Pa. L. Rev. 1495, 1498 (2013) (differentiating “software that serves as a ‘speech product,’” which should be subject to First Amendment protection, “from that which is a ‘communication tool,’” which should not receive such protections); Margot E. Kaminski, Privacy and the Right to Record, 97 B.U. L. Rev. 167, 172 (2017) (reconciling the “right to record” with “current First Amendment doctrine”).
See, e.g., Toni M. Massaro & Helen Norton, Siri-ously? Free Speech Rights and Artificial Intelligence, 110 Nw. U. L. Rev. 1169, 1173-75 (2016) (exploring how First Amendment theory and doctrine both leave room for protecting nonhuman speech); Toni M. Massaro, Helen Norton & Margot E. Kaminski, SIRI-OUSLY 2.0: What Artificial Intelligence Reveals About the First Amendment, 101 Minn. L. Rev. 2481, 2488-91 (2017) [hereinafter Massaro et al., SIRI-OUSLY 2.0] (explaining that the First Amendment may protect “strong AI speech” but that this need not deprive free speech of a human-centric focus).
See David S. Schwartz, Making Sense of 303 Creative: A Free Speech Solution in Search of a Problem 2-3 (Univ. Wis. L. Sch. Rsch. Paper Series, Paper No. 1792, 2024), https://ssrn.com/abstract=4702932 [https://perma.cc/J6JH-JVUQ] (arguing that the 303 Creative Court “blunders into [the] problem” of whether public-accommodations laws can “command compliance that is expressive of ideas that the regulated party disagrees with,” and addresses it partially, myopically, and badly). But see David D. Cole, “We Do No Such Thing”: 303 Creative v. Elenis and the Future of First Amendment Challenges to Public Accommodations Laws, 133 Yale L.J.F. 499, 501 (2024) (attempting to cabin the reach of 303 Creative by taking the Court “at its word” and restricting its reach to situations in which a commercial actor would refuse to serve all customers who seek the actor’s services to express a point of view with which the actor disagrees, as opposed to refusing to serve based on the status of the customer).