Stuck in the middleware with you: the challenges of capitalizing a market-oriented approach to platform governance
Open accessAbstract
Social media faces a crisis of trust over the management of speech, with critics simultaneously arguing platforms remove too much and too little. Middleware, a recent market-oriented proposal for platform governance, promises a way out of the impasse through third-party services for curation and content moderation. While large commercial actors like Meta and ByteDance have shown little interest in fostering competition, the situation is different for newer, more decentralized platforms. We approached Bluesky, which has substantial infrastructural support for third-party services, as a critical case to assess the potential of middleware in practice, investigating the implementation and capitalization of editorial services on the platform. Employing both quantitative and qualitative methods, we analyzed the uptake of Bluesky’s primary tools for curation (feeds, starter packs) and moderation (moderation lists, labels). On the supply side, we found evidence of significant interest, especially for curatorial services catering to online subcultures and left-leaning Bluesky users. Yet evidence of user demand was quite minimal, with low rates of adoption not indicative of a thriving marketplace. Our analysis offers mixed lessons for the future of middleware: if it is meant to function as a conventional market, substantial developments in monetization are necessary. If it is to be supported by diverse kinds of capital, the community infrastructure needed to support these endeavors is notoriously difficult to scale. Middleware as a marketplace of editorial services is thus not only a policy solution, but also a cultural contestation over the values of platform governance.
AI snapshot
Written from the paper’s full text.
TL;DR
Analyzing Bluesky as a critical case, this paper finds substantial supply of third-party curation and moderation services but minimal user adoption, undercutting middleware’s promise as a market-oriented platform-governance fix. It argues that middleware requires either viable monetization or hard-to-scale community infrastructure, making it a cultural contestation over platform values.
Key findings
- Bluesky offers a large middleware marketplace—nearly 80,000 feeds, 275,000 starter packs, over 14,000 moderation lists, and 362 custom labelers—but uptake is low: top-100 feeds averaged over 6,500 likes, with the 100th at 1,200; top-100 starter packs averaged 240 uses, with the 100th at 34; top-50 labelers averaged around 2,300 users, with the 50th at 173. Prior work cited by the authors found 9.4% of feeds never curated posts, 21.8% seemed inactive, and 4.37% of users had used a starter pack’s “follow all” operation.
- Curatorial supply skews English and toward identity, subculture, and left-leaning communities: identity-focused feeds were most prominent among top feeds; 92 of the top 100 starter packs were English; top starter packs focused on tech, news, and politics, with notable absences of right-leaning politics/news and mainstream entertainment, beauty, gaming, and food.
- Moderation lists are large and contentious: the top 100 largest lists averaged over 29,000 accounts, the biggest had over 180,000, 22 could not be loaded, and 13 referenced an appeals process; 90 of the top 100 were English. The most blocked accounts included Jesse Singal (79,562 blocks), Brianna Wu (60,183), Now Breezing (55,865), Felipe Neto (44,159), and Mark Cuban (43,554).
- Capitalization is mostly non-financial: the market functions largely as a gift economy sustained by subcultural capital, with minority experiments in tipping, subscriptions, ads, donations, and grants; Blacksky operated with ten paid employees funded by grants, seed-funding, and user donations, and Bluesky itself did not offer built-in tools for middleware vendors to monetize.
Methods
- The authors treat Bluesky as a critical case study and combine quantitative and qualitative digital methods, collecting 49 official Bluesky blog entries, 6 tutorials, 7 platform policies, and 50 interviews/public commentary items from tech publications.
- They analyzed four third-party features—feeds, starter packs, moderation lists, and labels—examining the top 100 iterations for the first three and the top 50 for labels due to lower uptake, coding source, primary language, and function; they also examined the 20 most blocked accounts and used Bluesky Directory and Clearsky for platform analytics.
- Starter packs were sampled by searching the English alphabet letter by letter, gathering the top 10 most used packs per letter, removing duplicates, and selecting the top 100; the authors also analyzed interviews, blog posts, and public statements from prominent actors leading third-party governance efforts.
Limitations
- The paper states that the starter-pack sampling method is an inexact approximation and likely underreports non-English packs; moderation-list use cannot be directly determined because lists can be used for public blocking or private muting, so the authors used list size instead; labels were limited to the top 50 due to low uptake; and the claim about partisan blocking orientation needs more research to confirm.
- The paper notes that data collection was completed before J.D. Vance joined Bluesky, and that it is ambiguous how Bluesky defines “use” for starter packs—whether it includes only new sign-ups, clicks on “follow all,” or follows of individual accounts.
- Evident from the text: the analysis relies on third-party analytics tools and platform metrics such as likes and reported uses, which the authors acknowledge do not directly translate to actual use; the Bluesky case is strategically chosen but may not generalize to large commercial platforms like Meta and ByteDance, which have shown little interest in middleware.
Model-written summary. Check it against the paper before relying on it.
Integrity signals
Registry
Cites
Statistics
Claims the paper makes
15 of 28 cited sources checked · 10 supported · 0 flagged · accuracy not yet measured
- Premisecitation
Social media platforms struggle to grapple with differences in perceptions of appropriate speech among transnational users.
“struggle to grapple with differences in perceptions of appropriate speech among transnational users (Jiang et al., 2021).”
Quote in the paperSubstantial claim1 of 1 sources checked- cites Jiang et al., 2021 source
- Premisecitation
The formerly open API on Twitter supported a network of third-party blocklists designed to help users manage harassment or curate their exposure to political speech.
“The formerly open API on the platform formerly known as Twitter used to support a network of third-party blocklists designed to, for example, help users manage harassment (Geiger, 2016) or curate their exposure to political speech (Wheatley & Vatnoey, 2020).”
Quote in the paperNot a substantial claim1 of 2 sources checked - Premisecitation
Regulatory solutions such as the European Union's Digital Services Act face significant jurisdictional hurdles in the United States and China, limiting the influence of the European Union's Brussels Effect.
“Yet such initiatives face significant jurisdictional hurdles in the United States and China (Liu & Yang, 2022), limiting the influence of the European Union's “Brussels Effect” (Li & Chen, 2024).”
Quote in the paperSubstantial claim1 of 2 sources checked- cites Liu & Yang, 2022 source
- cites Li & Chen, 2024 source
Section 2.3 (pp. 8–9) argues that the Brussels Effect has limitations in China, supporting the cited point that EU regulatory influence can be limited. But Li and Chen examine the GDPR and Chinese data-protection law; they do not establish that jurisdictional hurdles for initiatives such as the Digital Services Act in both the United States and China limit that influence. (LLM, conf=0.84)
Experimental — accuracy not yet measured
- Interpretationcitation
Some of the need for justifications may come from the fact that all moderation lists are public and might lead to a loss of face for those included.
“Some of the need for justifications may come from the fact that all moderation lists are, by definition, public and might lead to a loss of face for those included in them (Lopez & Ovaska, 2013).”
Quote in the paperNot a substantial claim1 of 1 sources checked- cites Lopez & Ovaska, 2013 source
Section 5.2 (page not available) says that increased visibility of unsocial actions can have undesirable interpersonal effects, and section 5.1 describes some such actions as face-threatening. The paper does not discuss moderation lists, establish that they are public, or link their publicity to a need for justifications or loss of face for people listed. (LLM, conf=0.82)
Experimental — accuracy not yet measured
- Interpretationcitation
Scaling middleware solutions requires a viable economic model.
“The situation leads us to agree that scaling middleware solutions requires a viable economic model (Hogg et al., 2024), a point we discuss below.”
Quote in the paperSubstantial claim1 of 1 sources checked- cites Hogg et al., 2024 source
- Findingown data
Twenty-two moderation lists could not be loaded.
“Still, we found plenty of evidence of maintenance problems: 22 moderation lists could not be loaded.”
Quote in the paperNot a substantial claim - Findingcitation
As of March 2025, 275,000 starter packs were on offer on Bluesky.
“While an impressive 275,000 starter packs were on offer as of March 2025,foot_5 their utility for recruitment and retention is hard to assess.”
Quote in the paperNot a substantial claim - Findingown data
Top starter packs varied in number of accounts from 8 to 242, averaging 74 accounts.
“Top starter packs varied in the number of accounts they contained, raining from 8 to 242, and averaging 74 accounts.”
Quote in the paperNot a substantial claim - Findingown data
English was the dominant language of 92 of the top 100 starter packs.
“Once again, English topped the list as the dominant language of 92 starter packs.”
Quote in the paperNot a substantial claim - Findingown data
Among the 100 most-liked feeds, the top two were official discovery feeds, each with over 35,000 likes.
“The top two were official discovery feeds, each with over 35,000 likes.”
Quote in the paperSubstantial claim - Findingown data
The average size of the top 100 moderation lists was over 29,000 accounts.
“The average size was over 29,000, an amount that almost certainly involves forms of automation to create and maintain.”
Quote in the paperNot a substantial claim - Findingown data
The average number of uses for the top 100 starter packs was 240.
“The average for the list was 240 uses.”
Quote in the paperNot a substantial claim - Findingnone given
Nearly 80,000 custom feeds are on offer on Bluesky.
“If Bluesky's custom feeds constitute a farmers' market of organic algorithmic choice, it's much bigger than your typical market, with nearly 80,000 feeds on offer.”
Quote in the paperNot a substantial claim - Findingown data
Thirteen moderation lists referenced an appeals process in the name or description.
“We also found some workarounds to make moderation lists more collaborative and publicly accountable, with 13 lists referencing an appeals process in the name or description.”
Quote in the paperNot a substantial claim - Findingown data
The biggest moderation list, Casey Ho's Platform Manipulation, Spam, & Coordinated Inauthentic Behavior, contained over 180,000 accounts.
“The biggest list, Casey Ho's “Platform Manipulation, Spam, & Coordinated Inauthentic Behavior, ” contained over 180,000 accounts, while the smallest, Sascha's “Andere Blocks (mod), ” contained over 10,000 accounts.”
Quote in the paperNot a substantial claim - Findingnone given
Politics accounts for the largest share of large moderation lists.
“While politics is one of the topics driving people to the platform, it also accounts for the largest share of large moderation lists.”
Quote in the paperNot a substantial claim - Findingcitation
Even Blacksky, one of the most developed services on the platform, operates with a team of ten paid employees, funded by a precarious combination of grants, seed-funding, and user donations.
“Even Blacksky, one of the most developed services on the platform, operates with a team of ten paid employees, funded by a precarious combination of grants, seed-funding, and user donations (Hendrix, 2025).”
Quote in the paperSubstantial claim0 of 1 sources checked- cites Hendrix, 2025
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
- Findingown data
On the supply side, there is evidence of significant interest, especially for curatorial services catering to online subcultures and left-leaning Bluesky users.
“On the supply side, we found evidence of significant interest, especially for curatorial services catering to online subcultures and left-leaning Bluesky users.”
Quote in the paperSubstantial claim - Findingnone given
Bluesky itself does not currently offer in-built tools for middleware vendors to monetize their offerings, highlighting an important limitation of the platform's infrastructural support.
“Notably, though, and as mentioned above, Bluesky itself does not currently offer in-built tools for middleware vendors to monetize their offerings, highlighting an important limitation of the platform's infrastructural support.”
Quote in the paperNot a substantial claim - Findingcitation
According to data from Bluesky Directory, there are 362 custom labelers.
“According to data from Bluesky Directory, there are 362 custom labelers.”
Quote in the paperNot a substantial claim - Findingown data
Our analysis of actually existing middleware offerings finds a conspicuous lack of financial capital.
“Our analysis of actually existing middleware offerings finds a conspicuous lack of financial capital.”
Quote in the paperSubstantial claim - Findingown data
Jesse Singal is the most blocked account on Bluesky, with nearly 80,000 blocks.
“Coming in first with nearly 80,000 blocks is Jesse Singal, an American journalist and podcaster who joined the platform in December 2024.”
Quote in the paperNot a substantial claim - Findingreasoning
While experiments with monetization are notable, they remain minority practices.
“While these experiments with monetization are notable, they remain minority practices.”
Quote in the paperSubstantial claim - Findingcitation
Most of the middleware market operates on subcultural capital-the community belonging that has long driven participation on the Web.
“Most of the market operates on a different kind of capital-the subcultural capital of community belonging that has long driven participation on the Web (Nissenbaum & Shifman, 2017).”
Quote in the paperSubstantial claim0 of 1 sources checked- cites Nissenbaum & Shifman, 2017 source
- Findingown data
Moderation lists showed a similar linguistic bias to custom feeds, with 90 of the top 100 in English.
“In terms of linguistic diversity, moderation lists showed a similar bias to the custom feeds, with 90 of the top 100 in English.”
Quote in the paperNot a substantial claim - Findingcitation
Much of the middleware market functions as a gift economy, with people making things to strengthen community ties.
“Much of the market functions as a gift economy, with people making things to strengthen community ties (Mauss, 2016).”
Quote in the paperSubstantial claim0 of 1 sources checked- cites Mauss, 2016
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
- Findingown data
Prominent moderation lists stand out for a thread of defensiveness across list descriptions.
“Prominent moderation lists stand out for the thread of defensiveness that cuts across list descriptions.”
Quote in the paperNot a substantial claim - Findingnone given
Inauthentic behavior is the next most prominent topic among large moderation lists, bringing together concern with spam, scams, engagement farming, and sock puppet accounts.
“Inauthentic behavior is the next most prominent topic, bringing together concern with spam, scams, engagement farming, and sock puppet accounts.”
Quote in the paperNot a substantial claim - Findingown data
The language breakdown of the 50 most followed labeling services closely mirrors the moderation list results, with 42 English-focused services.
“The language breakdown closely mirrors the moderation list results, with 42 English-focused services.”
Quote in the paperNot a substantial claim - Interpretationreasoning
Bluesky offers an ideal environment to observe evidence of supplier interest, service innovation, and customer demand.
“Accordingly, Bluesky offers an ideal environment to observe evidence of supplier interest, service innovation, and customer demand.”
Quote in the paperSubstantial claim - Premisenone given
Looking at newer and more decentralized platforms reveals substantially more compatibility with the premise of middleware.
“Looking at newer and more de-centralized platforms reveals substantially more compatibility with the premise of middleware.”
Quote in the paperSubstantial claim - Interpretationreasoning
The relatively low engagement metrics of top feeds suggest that there are many vendors with very little click traffic.
“Although likes don't directly translate to use, these relatively low engagement metrics suggest that there are a lot of vendors with very little click traffic.”
Quote in the paperNot a substantial claim - Interpretationown data
The identities of feed creators suggest a centralized marketplace, with Bluesky or its employees responsible for the largest share of top feeds.
“The identities of feed creators suggest a centralized marketplace, with Bluesky or its employees responsible for the largest share of top feeds.”
Quote in the paperNot a substantial claim - Interpretationreasoning
There seems to be substantially more interest in making feeds than in making use of them.
“To put it differently, there seems to be substantially more interest in making feeds than in making use of them.”
Quote in the paperSubstantial claim - Interpretationreasoning
This limitation makes it difficult to say whether the low levels of financial capitalization are due to technical challenges, a lack of market interest, or a combination of the two.
“This limitation makes it difficult to say whether the low levels of financial capitalization are due to technical challenges, a lack of market interest, or a combination of the two.”
Quote in the paperNot a substantial claim - Interpretationreasoning
Each community provides cultural infrastructure that complements the platform's technical infrastructure, offering social incentives for participation in a market where financial support is generally absent.
“Each community provides cultural infrastructure that complements the platform's technical infrastructure, offering social incentives for participation in a market where financial support is generally absent.”
Quote in the paperSubstantial claim - Interpretationown data
If middleware is meant to function as a conventional market, substantial developments in monetization are necessary to expand its offering, especially into more technically challenging and compute-intensive domains of video content.
“If it is meant to function as a conventional market, substantial developments in monetization are necessary to expand middleware's offering, especially into more technically challenging and compute-intensive domains of video content.”
Quote in the paperSubstantial claim - Interpretationown data
Our analysis highlights gaps between the promise and practice of middleware markets, organized around competing ideas of what the market is and who it serves.
“However, our analysis highlights some gaps between the promise and practice of middleware markets, organized around competing ideas of what the market is and who it serves.”
Quote in the paperSubstantial claim - Interpretationreasoning
The objection to Jesse Singal is less about personal exposure and more about a normative vision of public space and collective platform governance.
“The objection is less about personal exposure to Jesse Signal-although it may also be that-and more about a normative vision of public space and collective platform governance.”
Quote in the paperSubstantial claim - Interpretationreasoning
User disgruntlement in the face of ample available third-party tools shows the limitations of individual-oriented solutions.
“Similarly, user disgruntlement in the face of ample available third-party tools (Jesse Singal's name appears in the names and descriptions of 31 lists, and there are even platform label tools to identify and block followers of Jesse Singal's account) shows the limitations of individual-oriented solutions.”
Quote in the paperSubstantial claim - Interpretationreasoning
If middleware is meant to function as a mixed market supported by different kinds of capital, the community infrastructure needed to support these endeavors is notoriously difficult to scale, especially beyond established subcultures and marginalized communities.
“If it is meant to function as a mixed market, supported by very different kinds of capital, the community infrastructure needed to support these endeavors is notoriously difficult to scale, especially beyond established subcultures and marginalized communities.”
Quote in the paperNot a substantial claim - Interpretationnone given
If middleware is meant to function as a conventional market, substantial developments in monetization are necessary.
“if it is meant to function as a conventional market, substantial developments in monetization are necessary.”
Quote in the paperNot a substantial claim - Interpretationreasoning
Bluesky shifts algorithmic governance from an infrastructural component-centralized, inescapable, determining-to somewhere between a consumer product and a mode of public expression.
“While algorithmic decision-making is ubiquitous on social media, Bluesky shifts algorithmic governance from an infrastructural component-centralized, inescapable, determining-to somewhere between a consumer product and a mode of public expression.”
Quote in the paperSubstantial claim - Interpretationown data
The relatively low uptake of third-party editorial services offers a caution against too much optimism around middleware's appeal.
“And indeed, the relatively low uptake of thirdparty editorial services offers a caution against too much optimism around middleware's appeal.”
Quote in the paperSubstantial claim - Interpretationreasoning
As an experiment with a middleware approach to platform governance, the lessons of Bluesky are mixed.
“As an experiment with a middleware approach to platform governance, the lessons of Bluesky are mixed.”
Quote in the paperSubstantial claim
- Classified as probably not a substantial claim. 26
- May combine two claims. 33
- The quote only partly states this claim. 3
Classified as probably not a substantial claim.
PremisecitationBluesky has been described by proponents as a noteworthy and ambitious experiment in using middleware to actively incorporate the user community, increasing their agency over their social environment.
“Bluesky, in particular, has been described by proponents as a “noteworthy and ambitious experiment in using middleware to actively incorporate the user community… increasing their agency over their social environment” (Hogg et al., 2024, p. 16).”
Quote in the paperSubstantial claim1 of 1 sources checked- cites Hogg et al., 2024 source
Probably not a substantial claim: may not pass S3 (load-bearing).
PremisecitationEarlier research indicated that three feed generator providers, including SkyFeed, accounted for over 95% of all feeds on Bluesky.
“earlier research indicating that three providers, including SkyFeed, accounted for over 95% of all feeds on the platform (Balduf et al., 2024).”
Quote in the paperNot a substantial claim1 of 1 sources checked- cites Balduf et al., 2024 source
May combine two claims.
PremisecitationThere is currently no viable business model for third-party service providers, although ad-revenue sharing, subscription models, and reverse metering have been proposed.
“At the practical level, there is currently no viable business model for third-party service providers, although ad-revenue sharing arrangements, subscription models, and reverse metering have been proposed (Hogg et al., 2024).”
Quote in the paperSubstantial claim1 of 1 sources checked- cites Hogg et al., 2024 source
The For-Profit Models section proposes ad-revenue sharing, subscriptions, and reverse metering for middleware providers. The Potential Business Models section says finding sustainable models is important, but it does not say that no viable business model currently exists; it also describes Block Party’s former freemium model. The claim’s categorical opening is therefore unsupported. (LLM, conf=0.87)
Experimental — accuracy not yet measured
May combine two claims.
PremisecitationContent moderation challenges are compounded by opaque communication, resulting in a crisis of trust.
“These challenges are compounded by opaque communication surrounding content moderation (Cotter, 2023), resulting in a crisis of trust (Zuckerman & Rajendra-Nicolucci, 2023).”
Quote in the paperNot a substantial claim2 of 2 sources checkedProbably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.
PremisecitationLivestreaming platforms like Twitch feature a third-party market for content moderation tools to help streamers manage chatting audiences, even as such tools risk redundancy from Twitch's in-house offerings.
“Similarly, livestreaming platforms like Twitch feature a third-party market for content moderation tools to help streamers manage their chatting audiences (Han et al., 2023), even as such tools risk redundancy from Twitch's in-house offerings (Partin, 2020).”
Quote in the paperNot a substantial claim2 of 2 sources checked- cites Han et al., 2023 source
- cites Partin, 2020 source
Part II (pp. 7–8) describes Twitch introducing its own donation-management service, Twitch Bits, and gaining an advantage over third-party donation tools. That supports the general risk posed by an in-house offering, but the paper does not establish that third-party content-moderation tools face that risk. The third-party moderation market is attributed to Han et al., not Partin. (LLM, conf=0.82)
Experimental — accuracy not yet measured
May combine two claims.
PremisecitationBluesky's experiments with governance are all about expanding individual choice, contributing to the platform's association with liberal values.
“Bluesky's experiments with governance are all about expanding individual choice, contributing to the platform's association with liberal values (Tkacz & Gehl, 2025).”
Quote in the paperNot a substantial claim1 of 1 sources checked- cites Tkacz & Gehl, 2025 source
The discussion of platform polities (p. 5) describes Bluesky as aiming to embody a ‘liberal internet’ aligned with liberal values. The discussion of Bluesky’s protocol (p. 3) describes decentralisation and the possibility of users’ ‘credible exit,’ but does not say its governance experiments are all about expanding individual choice. The same platform-polities section (p. 5) cautions that choice alone cannot explain differences between platforms. (LLM, conf=0.86)
Experimental — accuracy not yet measured
Probably not a substantial claim: may not pass S3 (load-bearing).
PremisecitationBalduf et al. concluded that 4.37% of users had employed the follow all operation on a starter pack.
“This aligns with Balduf et al. 's conclusion that “4.37% [of ] users had employed the follow all operation on a starter pack” (2025, p. 4, emphasis in original).”
Quote in the paperNot a substantial claim1 of 1 sources checked- cites Balduf et al., 2025 source
May combine two claims.
PremisecitationA recent analysis found that 9.4% of all feeds had never curated any posts while another 21.8% seemed to be inactive.
“a recent analysis found that 9.4% of all feeds had never curated any posts while another 21.8% seemed to be inactive (Balduf et al., 2024).”
Quote in the paperSubstantial claim1 of 1 sources checked- cites Balduf et al., 2024 source
Probably not a substantial claim: may not pass S3 (load-bearing).
Findingnone givenskyfeed.xyz had four feeds for managing spam in the top 100.
“The final top user is skyfeed.xyz,with four feeds for managing spam.”
Quote in the paperNot a substantial claimMay combine two claims.
Findingown dataAmong the 100 most-liked feeds, popularity metrics plummet down the list, with only 1,200 likes at the bottom for the Japanese-language feed JP, and the list averages over 6,500 likes.
“The popularity metrics plummet as we move down the list, with only 1,200 likes at the bottom for “JP, ” a feed that aggregates Japanese language posts.The list averages over 6,500 likes.”
Quote in the paperSubstantial claimMay combine two claims.
Findingown dataIdentity-focused feeds were most prominent among top feeds, serving national, racial, sexual, gender, and disability affiliations; discovery was next most popular; art, relationships, NSFW, books, news, science, sports, and furries each had at least five feeds in the top 100.
“In terms of topic, identity-focused feeds were most prominent, serving different national, racial, sexual, and gender affiliations, along with the disability community.Discovery was the next most popular function and the following categories had at least five feeds within the top 100: art, relationships, NSFW, books, news, science, sports, and furries.”
Quote in the paperNot a substantial claimMay combine two claims. The quote only partly states this claim.
Findingown dataTwo users, clarabelle.xyz and furryli.st, each had five feeds in the top 100.
“Two users, clarabelle.xyz and furryli.st, each had five feeds.”
Quote in the paperNot a substantial claimProbably not a substantial claim: may not pass S3 (load-bearing).
Findingown dataFour moderation lists mentioned using the same online tool, Blockenheimer, in their descriptions.
“Indeed, four moderation lists mentioned using the same online tool-Blockenheimer-in their descriptions.”
Quote in the paperNot a substantial claimMay combine two claims.
Findingown dataTop feeds heavily skew towards English content, with five Japanese, four Portuguese, and one each Indonesian, Ukrainian, and Spanish.
“Top feeds heavily skew towards English content, although there were five Japanese, four Portuguese, and a single feed each for Indonesian, Ukrainian, and Spanish.”
Quote in the paperNot a substantial claimProbably not a substantial claim: may not pass S3 (load-bearing).
Findingnone givenDestiny is the subject of four starter packs.
“The subject of four starter packs, Destiny is one of the most senior livestreamers who began as a Starcraft II streamer in the days of Justin.TV and Ustream.”
Quote in the paperNot a substantial claimMay combine two claims.
Findingown dataJaz, a self-described gender nomad and former Bluesky backend developer, had eight feeds in the top 100.
“The next most prominent creator was Jaz, a self-described “gender nomad” 11 with eight feeds in the top 100.”
Quote in the paperNot a substantial claimMay combine two claims.
FindingcitationThere were notable absences in top starter packs, including a general lack of right-leaning politics or news groups, mainstream entertainment media apart from Taylor Swift fans, and other prominent social media topics like beauty, gaming, and food.
“There were some notable absences, including a general lack of right-leaning politics or news groups, mainstream entertainment media (apart from Taylor Swift fans), and other prominent social media topics like beauty, gaming, and food (Hallinan et al., 2023).”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites Hallinan et al., 2023 source
May combine two claims.
Findingown dataThe biggest categories of top starter packs focus on tech, news, and politics, including the DGGverse for fans of Destiny.
“The biggest categories focus on tech, news, and politics, including the DGGverse for fans of American political commentator and livestreamer Destiny”
Quote in the paperNot a substantial claimMay combine two claims.
Findingown dataStarter packs are difficult to locate directly within the app; they do not appear in search results and only come up if people post about one or visit a creator's profile.
“However, starter packs are difficult to locate directly within the app.They don't appear in search results and instead only come up if people post about a starter pack or if a user visits the profile of someone who has created one (see Fig. 1).”
Quote in the paperSubstantial claimMay combine two claims.
Findingnone givenBluesky users have created over 14,000 moderation lists, roughly one-sixth the total number of custom feeds and one-twentieth the number of starter packs.
“Bluesky users have created over 14,000 moderation lists, roughly one-sixth of the total number of custom feeds and one-twentieth the number of starter packs on the platform.”
Quote in the paperNot a substantial claimProbably not a substantial claim: may not pass S3 (load-bearing).
Findingown dataIn one hundredth place among top starter packs was the Flesh and Blood TCG Starter Pack, used only 34 times.
“In one hundredth place was the “Flesh and Blood TCG Starter Pack” for fans of the trading card game, which had been used only 34 times.”
Quote in the paperNot a substantial claimProbably not a substantial claim: may not pass S4 (based). May combine two claims.
Findingnone givenBluesky promises and delivers a marketplace of editorial services, with hundreds of thousands of third-party options available for users to customize their social media experience.
“Bluesky promises and delivers a marketplace of editorial services, with hundreds of thousands of third-party options available for users to customize their social media experience.”
Quote in the paperNot a substantial claimMay combine two claims.
Findingnone givenDespite Jesse Singal being the most blocked account on Bluesky (by 80,000 accounts at the time of data collection), users circulated an online petition with over 26,000 signatures calling for his ban.
“The case of Jesse Singal highlights this dispute: despite being the most blocked account on the platform (by 80,000 accounts at the time of data collection), users still circulated an online petition with over 26,000 signatures calling for his ban.”
Quote in the paperNot a substantial claimMay combine two claims.
Findingown dataEvidence of user demand was quite minimal, with low rates of adoption not indicative of a thriving marketplace.
“Yet evidence of user demand was quite minimal, with low rates of adoption not indicative of a thriving marketplace.”
Quote in the paperSubstantial claimProbably not a substantial claim: may not pass S3 (load-bearing), S4 (based).
Findingnone givenDiscovery and interest-based offerings serve individual preferences above all.
“Discovery and interest-based offerings serve individual preferences above all.”
Quote in the paperNot a substantial claimMay combine two claims.
Findingnone givenThe connections to online subcultures in Bluesky middleware are abundant and include furries, fans, artists, sex workers, and political streamers, as well as substantial representation of social minorities such as LGBTQ communities, disability communities, and Blacksky.
“The connections to online subcultures are abundant and include the impressive organization of furries, fans, artists (especially of the lowbrow variety), sex workers, and political streamers (reflected in the distinctive if polarizing popularity of Destiny).There is also substantial representation of social minorities, which the examples of LGBTQ communities, disability communities, and Blacksky”
Quote in the paperNot a substantial claimMay combine two claims.
Findingown dataCross-referencing large moderation lists with the most blocked accounts on Bluesky suggests a partisan orientation for blocking on the platform, with conservative accounts blocked more frequently than liberals, although more research is needed to confirm this claim.
“Cross-referencing large moderation lists with the most blocked accounts on Bluesky suggests a partisan orientation for blocking on the platform, with conservative accounts blocked more frequently than liberals, although more research is needed to confirm this claim.”
Quote in the paperNot a substantial claimMay combine two claims.
Findingnone givenMost political block lists target the right or resist liberals that oppose U.S. President Donald Trump.
“However, most of the political block lists target the right (e.g., “Conservative-Following Blocklist (Appeals open for progressives)”) or “resist” liberals that oppose U.S. President Donald Trump (e.g., “#Resistance Dweebs”).”
Quote in the paperNot a substantial claimProbably not a substantial claim: may not pass S4 (based).
Findingnone givenFor most of the popular feeds, starter packs, moderation lists, and labeler services, the connection to truth is less obvious.
“However, for most of the popular feeds, starter packs, moderation lists, and labeler services, the connection to truth is less obvious.”
Quote in the paperNot a substantial claimProbably not a substantial claim: may not pass S3 (load-bearing), S4 (based).
FindingreasoningBadge labels are a more popular application of labeling services, perhaps because they require less effort and can be configured as a self-service model.
“Badge labels are a more popular application of the service, perhaps because they require less effort and can be configured as a self-service model.”
Quote in the paperNot a substantial claimMay combine two claims. The quote only partly states this claim.
Findingown dataThe 50 most followed labeling services on Bluesky have an average of around 2,300 users, with the 50th-most followed account having only 173 users.
“The average number of users for the services we looked at was around 2,300 users.”
Quote in the paperSubstantial claimProbably not a substantial claim: may not pass S3 (load-bearing), S4 (based). May combine two claims.
Findingnone givenUnlike custom feeds and starter packs, the offerings of labeling services are not extensive, but like them, only a very small minority of Bluesky users appear to have made use of this middleware feature.
“Unlike custom feeds and starter packs, the offerings here are not extensive, but like them, only a very small minority of Bluesky users appear to have made use of this middleware feature.”
Quote in the paperNot a substantial claimClassified as probably not a substantial claim.
PremisereasoningRather than counter the power of commercial platforms through regulation, middleware seeks to empower the user through increased editorial choice.
“Rather than counter the power of commercial platforms through regulation, middleware seeks to empower the user through increased editorial choice.”
Quote in the paperSubstantial claimMay combine two claims.
PremisecitationPrevious research on Twitter blocklists found that some users who had been added to blocklists did not understand why they were there, were not able to change their status, and suffered reputational harms.
“previous research on Twitter blocklists found that some users who had been added to blocklists didn't understand why they were there, were not able to change their status, and suffered reputational harms (Jhaver et al., 2018;Wheatley & Vatnoey, 2020)”
Quote in the paperSubstantial claim0 of 2 sources checkedClassified as probably not a substantial claim. May combine two claims.
PremisecitationMiddleware is a recent market-oriented proposal for platform governance that promises a way out of the impasse through third-party services for curation and content moderation.
“Middleware, a recent proposal for a market-oriented approach to platform governance, promises a way out of the impasse through third-party services for curation and content moderation.”
Quote in the paperSubstantial claimProbably not a substantial claim: may not pass S4 (based).
Premisenone givenBluesky's design and corporate ethos make it especially easy for third parties, including individuals, community groups, non-profit organizations, and aspiring businesses, to make their own editorial services and make these services available to users.
“Basically, the design and corporate ethos of Bluesky make it especially easy for third parties, including individuals, community groups, non-profit organizations, and aspiring businesses, to make their own editorial services and make these services available to users.”
Quote in the paperNot a substantial claimMay combine two claims.
InterpretationreasoningStarter packs are primarily intended for external recruitment, circulating in other information environments and directing people back to Bluesky.
“As these design choices suggest, starter packs are primarily intended for external recruitment, circulating in other information environments and then directing people back to Bluesky.”
Quote in the paperNot a substantial claimClassified as probably not a substantial claim.
PremisereasoningBy introducing market competition at the editorial level, middleware aims to reduce the influence of large commercial platforms over public life.
“By introducing market competition at the editorial level, middleware aims to reduce the influence of large commercial platforms over public life.”
Quote in the paperSubstantial claimMay combine two claims.
Interpretationown dataAs with custom feeds, there is a great number of starter packs on offer but little evidence of use.
“As with custom feeds, here too we found a great number of starter packs on offer, but little evidence of use.”
Quote in the paperSubstantial claimMay combine two claims.
PremisecitationEncouraging customized social media experiences is unlikely to address social fragmentation and ideological polarization, given that it does not change the underlying ad-driven business model.
“Furthermore, encouraging customized social media experiences is unlikely to address social fragmentation and ideological polarization (Faris & Donovan, 2021), given that it does not change the underlying ad-driven business model (Maréchal, 2021).”
Quote in the paperSubstantial claim0 of 2 sources checkedClassified as probably not a substantial claim.
PremisecitationBluesky's approach to curation is based on the idea that people must have control over algorithms to trust online spaces.
“Bluesky's approach to curation is based on the idea that people “must have control over our algorithms if we're going to trust in our online spaces” (The Bluesky Team, 2022).”
Quote in the paperSubstantial claim0 of 1 sources checked- cites The Bluesky Team, 2022
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
The quote only partly states this claim.
Study design and conductreasoningBecause moderation lists can be used for both public blocking and private muting, the paper cannot directly determine which lists are used the most.
“we cannot directly determine which lists are used the most.”
Quote in the paperSubstantial claimProbably not a substantial claim: may not pass S4 (based).
Premisenone givenDespite widespread support among public intellectuals and Trust & Safety professionals, real world implementation of editorial middleware remains relatively limited.
“Despite widespread support among public intellectuals and Trust & Safety professionals, real world implementation of editorial middleware remains relatively limited.”
Quote in the paperNot a substantial claim- Rests on an assumption: Large commercial actors like Meta and ByteDance have shown little interest in opening their platforms to competition without a regulatory stick and/or carrot.
May combine two claims.
Interpretationown dataOverall, there is a large number of feed products but a very long tail of user uptake, suggesting an imbalance between supply and demand.
“Overall, we see a large number of products, but a very long tail of user uptake, suggesting an imbalance between supply and demand.”
Quote in the paperSubstantial claimProbably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing).
PremisereasoningThere is a need for regulation to provide and preserve access to the APIs of commercial platforms.
“There is also the need for regulation to provide and preserve access to the APIs of commercial platforms, especially given trends toward greater restrictions from major actors like Meta, X, and Reddit.”
Quote in the paperNot a substantial claimClassified as probably not a substantial claim.
PremisereasoningMiddleware reflects a distinctly market- and free speech-oriented approach to platform governance that seeks minimal state intervention.
“Unsurprisingly, given its pedigree, middleware reflects a distinctly market-and free speech-oriented approach to platform governance that seeks minimal state intervention.”
Quote in the paperSubstantial claimProbably not a substantial claim: may not pass S3 (load-bearing), S4 (based).
Interpretationnone givenAt a technical level, Bluesky's design is largely a success, offering a social media experience that replicates the feel of centralized social media platforms.
“At a technical level, the design is largely a success, offering a social media experience that replicates the feel of centralized social media platforms.”
Quote in the paperNot a substantial claimProbably not a substantial claim: may not pass S2 (checkable), S4 (based). May combine two claims.
InterpretationreasoningMiddleware as a marketplace of editorial services is not only a policy solution but also a cultural contestation over the values of platform governance.
“Middleware as a marketplace of editorial services is thus not only a policy solution, but also a cultural contestation over the values of platform governance.”
Quote in the paperNot a substantial claimMay combine two claims.
InterpretationreasoningThere is no hard boundary between financial and subcultural capital; users may indirectly monetize the attention they receive for creating community tools.
“Of course, there is no hard boundary between financial and subcultural capital; users may indirectly monetize the attention they receive for creating community tools.”
Quote in the paperNot a substantial claimProbably not a substantial claim: may not pass S2 (checkable), S4 (based).
Interpretationnone givenIf middleware is supported by diverse kinds of capital, the community infrastructure needed to support these endeavors is notoriously difficult to scale.
“If it is to be supported by diverse kinds of capital, the community infrastructure needed to support these endeavors is notoriously difficult to scale.”
Quote in the paperNot a substantial claimMay combine two claims.
Interpretationown dataWe already see friction between Bluesky's leadership and some of its users that suggest competing understandings of the platform, most evident in disputes over top-level content moderation.
“While cultural practices and ambiguous engagement metrics help sustain multiple and potentially competing ideas of value (Hallinan & Brubaker, 2021), we already see friction between Bluesky's leadership and some of its users that suggest competing understandings of the platform.This conflict is most evident in disputes over top-level content moderation.”
Quote in the paperSubstantial claim0 of 1 sources checked- cites Hallinan & Brubaker, 2021
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
May combine two claims.
Study design and conductreasoningThe starter pack sample is an inexact approximation that likely underreports non-English packs because it was gathered by searching the English alphabet letter by letter and selecting the top 100 by use.
“The results are thus an inexact approximation, but still give a good sense of the attributes of popular starter packs on the platform.It likely underreports non-English packs given the sampling method.”
Quote in the paperNot a substantial claimProbably not a substantial claim: may not pass S3 (load-bearing).
PremisecitationJesse Singal has been cataloged by GLAAD's Accountability Project for his writings on transgender issues and other matters.
“As reporting from TechCrunch notes, Singal “has been cataloged by GLAAD's Accountability Project for his writings on transgender issues and other matters””
Quote in the paperNot a substantial claim0 of 1 sources checked- cites TechCrunch
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
May combine two claims.
InterpretationreasoningThe openness and customizability of Bluesky support substantial innovation, and the rise of feed generators-as-a-service further expands participation by lowering the technical skills and resources necessary for people to create and share their own algorithms.
“At the same time, the openness and customizability of the platform support substantial innovation.The rise of feed generators-as-a-service further expands participation, lowering the technical skills and resources necessary for people to create (and share) their own algorithms.”
Quote in the paperSubstantial claimClassified as probably not a substantial claim.
Study design and conductnone givenIt is ambiguous how Bluesky defines use for starter packs.
“It is, however, ambiguous how Bluesky defines use.”
Quote in the paperSubstantial claim
The paper relies on this without support.
Large commercial actors like Meta and ByteDance have shown little interest in opening their platforms to competition without a regulatory stick and/or carrot.
“Large commercial actors like Meta and ByteDance have shown little interest in opening their platforms to competition without a regulatory stick and/or carrot, although there are notable exceptions.”
The paper relies on this without support.
Third-party offerings for content recommendation are quite marginal in the historical record.
“AThird-party offerings for content recommendation, however, are quite marginal in the historical record.”
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
Each contradiction between two passages listed here was confirmed by an independent check.
Starter pack exceeds the stated account limit
MinorThe paper says a starter pack can contain up to 150 users, but later reports that a top starter pack contained 242 accounts.
“Users can also actively curate their experience of the app with starter packs: lists containing up to 150 users and three feeds to follow (The Bluesky Team, 2024b).”
“Top starter packs varied in the number of accounts they contained, raining from 8 to 242, and averaging 74 accounts.”
Cites
Five misunderstandings about case-study research
Flyvbjerg, B. (2006) · Qualitative Inquiry · doi:10.1177/1077800405284363 · Cited by 14,266 · cited 2× · see citing works
“Building on this observation, we approach Bluesky as a “critical case” to assess the potential of middleware in practice (Flyvbjerg, 2006, p. 229), investigating the implementation and capitalization of third-party editorial services on the platform.”
Toward a Comparative Sociology of Valuation and Evaluation
Lamont, M. (2012) · Annual Review of Sociology · doi:10.1146/annurev-soc-070308-120022 · Cited by 1,492 · cited 2× · see citing works
“We then provide background context on Bluesky and introduce the sociology of valuation as our theoretical framework (Lamont, 2012), moving beyond a narrow economic understanding of value to understand the diverse markets fostered by social media (Tkacz & Gehl, 2025).”
Identity and deception in the virtual community
Donath, J. (2002) · Routledge eBooks · doi:10.4324/9780203194959-11 · Cited by 1,177 · cited 1× · see citing works
“Online communities have long employed intermediary governance solutions, exemplified by the development of killfiles on Usenet, referring to shared filters to avoid particular people or topics (Donath, 2002).”
Recommended for you: The Netflix Prize and the production of algorithmic culture
Hallinan, B., & Striphas, T. (2014) · New Media & Society · doi:10.1177/1461444814538646 · Cited by 542 · see citing works
Self-citationWhat is a flag for? Social media reporting tools and the vocabulary of complaint
Crawford, K., & Gillespie, T. (2016) · New Media & Society · doi:10.1177/1461444814543163 · Cited by 496 · cited 1× · see citing works
“…amount of individual labor required to configure social media to one’s liking and highlight competing approaches to content moderation, potentially prompting debates about platforms and the public sphere in a way that centralized user reporting deliberately obfuscates (Crawford & Gillespie, 2016).”
Governing online platforms: From contested to cooperative responsibility
Helberger, N., Pierson, J., & Poell, T. (2017) · The Information Society · doi:10.1080/01972243.2017.1391913 · Cited by 423 · see citing works
Internet memes as contested cultural capital: The case of 4chan’s /b/ board
Nissenbaum, A., & Shifman, L. (2017) · New Media & Society · doi:10.1177/1461444815609313 · Cited by 418 · see citing works
Affordances-in-practice: An ethnographic critique of social media logic and context collapse
Costa, E. (2018) · New Media & Society · doi:10.1177/1461444818756290 · Cited by 363 · cited 1× · see citing works
“However, the existence of a feature can only tell us so much about its use — a point Elisabetta Costa forcefully makes with the notion of “affordances-in-practice” (2018).”
Katzenbach, C., & Ulbricht, L. (2019) · Internet Policy Review · doi:10.14763/2019.4.1424 · Cited by 359 · see citing works
Online Harassment and Content Moderation: The Case of Blocklists
Jhaver, S., Ghoshal, S., Bruckman, A., & Gilbert, E. (2018) · ACM Transactions on Computer-Human Interaction · doi:10.1145/3185593 · Cited by 270 · cited 4× · see citing works
“…found that some users who had been added to blocklists didn’t understand why they were there, were not able to change their status, and suffered reputational harms (Jhaver et al., 2018; Wheatley & Vatnoey, 2020), and replies to an official Bluesky post about blocklists show similar concerns.”
InfluentialBrock, A. (2020) · doi:10.18574/9781479811908 · Cited by 235 · see citing works
Cotter, K. (2023) · Information, Communication & Society · doi:10.1080/1369118x.2021.1994624 · Cited by 149 · cited 1× · see citing works
“These challenges are compounded by opaque communication surrounding content moderation (Cotter, 2023), resulting in a crisis of trust (Zuckerman & Rajendra-Nicolucci, 2023).”
Geiger, R. S. (2016) · Information, Communication & Society · doi:10.1080/1369118x.2016.1153700 · Cited by 125 · cited 4× · see citing works
“In so doing, middleware would help users address the “per-ceived governance gap” (Geiger, 2016, p. 789) between a platform’s definition of permissible behavior and their own understanding.”
InfluentialA Pragmatic Redefinition of Value(s): Toward a General Model of Valuation
Heinich, N. (2020) · Theory Culture & Society · doi:10.1177/0263276420915993 · Cited by 108 · cited 4× · see citing works
“The “actual forms under which valuation occurs” includes mechanisms of price, social attachment, and public judgment (Heinich, 2020, p. 83), meaning that estimates of economic worth represent one among many possible measures of value.”
InfluentialBit by (Twitch) Bit: “Platform Capture” and the Evolution of Digital Platforms
Partin, W. C. (2020) · Social Media + Society · doi:10.1177/2056305120933981 · Cited by 79 · cited 1× · see citing works
“Similarly, livestreaming platforms like Twitch feature a third-party market for content moderation tools to help streamers manage their chatting audiences (Han et al., 2023), even as such tools risk redundancy from Twitch’s in-house offerings (Partin, 2020).”
Managing complexity: middleware explained
Campbell, A. T., Coulson, G., & Kounavis, M. E. (1999) · IT Professional · doi:10.1109/6294.793667 · Cited by 69 · cited 1× · see citing works
“The name of the proposal riffs on the longstanding use of middleware in software development, where it refers to a software layer that operates between system and application (Campbell et al., 1999).”
Understanding international perceptions of the severity of harmful content online
Jiang, J. A., Scheuerman, M. K., Fiesler, C., & Brubaker, J. R. (2021) · PLOS ONE · doi:10.1371/journal.pone.0256762 · Cited by 69 · cited 1× · see citing works
“struggle to grapple with differences in perceptions of appropriate speech among trans-national users (Jiang et al., 2021).”
“Dual‐Track” platform governance on content: A comparative study between China and United States
Liu, J., & Yang, L. (2022) · Policy & Internet · doi:10.1002/poi3.307 · Cited by 48 · cited 1× · see citing works
“Yet such initiatives face significant jurisdictional hurdles in the United States and China (Liu & Yang, 2022), limiting the influence of the European Union’s “Brussels Effect” (Li & Chen, 2024).”
Hate Raids on Twitch: Echoes of the Past, New Modalities, and Implications for Platform Governance
Han, C., Seering, J., Kumar, D., Hancock, J. T., & Durumeric, Z. (2023) · Proceedings of the ACM on Human-Computer Interaction · doi:10.1145/3579609 · Cited by 46 · cited 2× · see citing works
“Similarly, livestreaming platforms like Twitch feature a third-party market for content moderation tools to help streamers manage their chatting audiences (Han et al., 2023), even as such tools risk redundancy from Twitch’s in-house offerings (Partin, 2020).”
Testimonial rallies and the construction of memetic authenticity
Shifman, L. (2018) · European Journal of Communication · doi:10.1177/0267323118760320 · Cited by 44 · see citing works
Rogers, R. A. (2024) · SAGE Publications Ltd eBooks · doi:10.4135/9781036231514 · Cited by 41 · cited 2× · see citing works
“We describe our digital methods approach to platform research (Rogers, 2024) and present our analysis of the primary third-party curation (feeds, starter packs) and moderation tools (moderation lists, Labels) on Bluesky.”
Distributed Blackness: African American Cybercultures
Brock, A. (2020) · doi:10.18574/nyu/9781479847228.001.0001 · Cited by 40 · see citing works
Li, W., & Chen, J. (2024) · Computer Law & Security Review · doi:10.1016/j.clsr.2024.105994 · Cited by 31 · cited 1× · see citing works
“Yet such initiatives face significant jurisdictional hurdles in the United States and China (Liu & Yang, 2022), limiting the influence of the European Union’s “Brussels Effect” (Li & Chen, 2024).”
Wheatley, D., & Vatnoey, E. (2020) · New Media & Society · doi:10.1177/1461444819858278 · Cited by 28 · cited 2× · see citing works
“…found that some users who had been added to blocklists didn’t understand why they were there, were not able to change their status, and suffered reputational harms (Jhaver et al., 2018; Wheatley & Vatnoey, 2020), and replies to an official Bluesky post about blocklists show similar concerns.”
Mapping the transnational imaginary of social media genres
Hallinan, B., Kim, B., Scharlach, R., Trillò, T., Mizoroki, S., & Shifman, L. (2023) · New Media & Society · doi:10.1177/14614448211012372 · Cited by 24 · see citing works
Self-citationGoverning trust in European platform societies: Introduction to the special issue
van Dijck, J. (2021) · European Journal of Communication · doi:10.1177/02673231211028378 · Cited by 22 · cited 1× · see citing works
“…some legislation may be necessary to open commercial platforms to third-party competition, the strategy’s emphasis on economic incentives and technological solutions stands in strong contrast to the emphasis on sovereignty and fundamental rights typical to European approaches (Van Dijck, 2021).”
Zuckerman, E., & Rajendra-Nicolucci, C. (2023) · Social Media + Society · doi:10.1177/20563051231196864 · Cited by 22 · cited 1× · see citing works
“These challenges are compounded by opaque communication surrounding content moderation (Cotter, 2023), resulting in a crisis of trust (Zuckerman & Rajendra-Nicolucci, 2023).”
Making the Internet Safe for Democracy
Fukuyama, F. (2021) · Journal of democracy · doi:10.1353/jod.2021.0017 · Cited by 19 · cited 1× · see citing works
“…professionals, and policy makers, best reflected in the publication of a special issue devoted to the topic in the Journal of Democracy in 2021 (Fukuyama, 2021) and a 2024 Symposium jointly hosted by the Foundation for American Innovation and Stanford’s Cyber Policy Center (Hogg et al., 2024).”
A Look at Unsociability on Facebook
Lopez, M. G., & Ovaska, S. (2013) · Electronic Workshops in Computing · doi:10.14236/ewic/hci2013.15 · Cited by 16 · cited 1× · see citing works
“Some of the need for justifications may come from the fact that all moderation lists are, by definition, public and might lead to a loss of face for those included in them (Lopez & Ovaska, 2013).”
Looking AT the Blue Skies of Bluesky
Balduf, L., Sokoto, S., Ascigil, O., Tyson, G., Scheuermann, B., Korczyński, M., Castro, I., & Król, M. (2024) · Proceedings of the 2024 ACM on Internet Measurement Conference · doi:10.1145/3646547.3688407 · Cited by 16 · cited 2× · see citing works
“However, the existence of a feed does not guarantee its operation; a recent analysis found that 9.4% of all feeds had never curated any posts while another 21.8% seemed to be inactive (Balduf et al., 2024).”
InfluentialThe Future of Platform Power: Fixing The Business Model
Maréchal, N. (2021) · Journal of Democracy · doi:10.1353/jod.2021.0041 · Cited by 12 · cited 1× · see citing works
“Furthermore, encouraging customized social media experiences is unlikely to address social fragmentation and ideological polarization (Faris & Donovan, 2021), given that it does not change the underlying ad-driven business model (Maréchal, 2021).”
The Future of Platform Power: Quarantining Misinformation
Faris, R., & Donovan, J. (2021) · Journal of Democracy · doi:10.1353/jod.2021.0040 · Cited by 7 · cited 1× · see citing works
“Furthermore, encouraging customized social media experiences is unlikely to address social fragmentation and ideological polarization (Faris & Donovan, 2021), given that it does not change the underlying ad-driven business model (Maréchal, 2021).”
Bootstrapping Social Networks: Lessons from Bluesky Starter Packs
Balduf, L., Sokoto, S., Baronchelli, A., Castro, I., Król, M., Tyson, G., Pavlou, G., Scheuermann, B., & Ascigil, O. (2025) · Proceedings of the International AAAI Conference on Web and Social Media · doi:10.1609/icwsm.v19i1.35810 · Cited by 6 · see citing works
After Twitter: Fragmentation, Platform Polities and Protective Sociality
Tkacz, N., & Gehl, R. W. (2025) · Social Media + Society · doi:10.1177/20563051251366907 · Cited by 5 · cited 4× · see citing works
“We then provide background context on Bluesky and introduce the sociology of valuation as our theoretical framework (Lamont, 2012), moving beyond a narrow economic understanding of value to understand the diverse markets fostered by social media (Tkacz & Gehl, 2025).”
InfluentialPriorities and exclusions within Trust and Safety industry standards
Hallinan, B., Reynolds, C., Scharlach, R., Theiler, D., Niv, N., Rothenstein, O., Knief, I., & Kuperberg, Y. (2025) · New Media & Society · doi:10.1177/14614448251357225 · Cited by 2 · cited 1× · see citing works
“Alternative social media platforms already offer a version of this, attracting users because of distinct ideological commitments woven into the public communication, design, and policies of particular platforms (Hallinan et al., 2025).”
Self-citationShaping the Future of Social Media with Middleware
Hogg, L., DiResta, R., Fukuyama, F., Reisman, R., Keller, D., Ovadya, A., Thorburn, L., Stray, J., & Mathur, S. (2024) · arXiv · doi:10.48550/arxiv.2412.10283
Flagged first, then by citations
Research works printed but not linked: 4
- Hallinan, B., & Brubaker, J. R. (2021). Living with everyday evaluations on social media platforms. International Journal of Communication, 15, 1551–1569.
- Hendrix, J. (2025). Podcast: Rudy Fraser on building Blacksky and the future of middleware. Tech Policy Press. https://www.techpolicy.press/podcast-rudy-fraser-on-building-blacksky-and-the-future-of-middleware/
- Lai, S., & Roth, Y. (2024). Online safety and the “great decentralization”—The perils and promises of federated social media. Tech Policy Press. https://www.techpolicy.press/online-safety-and-the-great-decentralization-the-perils-and-promises-of-federated-social-media/
- Mauss, M. (2016). The Gift (J. I. Guyer, Trans.; Expanded edition). HAU Books.
- Authors
- Hallinan, B., Rothenstein, O., & John, N. A.
- Affiliations
- Hebrew University of Jerusalem (IL)
- Corresponding author
- Hallinan, Blake
- Venue
- Communication and Change
- Published
- 2025-12-01
- Volume
- 1
- Issue
- 1
- ISSN
- 3059-2011 (ISSN-L)
- OpenAlex
- W4416832765
- Cited by
- 0
- References
- 36
- FWCI
- 0.00
- Type
- journal article
- Language
- en
- Open access
- hybrid
- doi.org
- Locations
- Communication and Change (publishedVersion, cc-by)
- Research Explorer (The University of Manchester) (submittedVersion, other-oa)
- Funders
- Israel Science Foundation (Award #3341/24)
- Topics
- Social Sciences › Communication › Social Media and Politics
- Social Sciences › Business, Management and Accounting › Strategy and Management › Digital Platforms and Economics
- Social Sciences › Law › Freedom of Expression and Defamation
- Keywords
- middleware, platform governance, social media, content moderation, content curation, editorial services, Bluesky, decentralized platforms
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Claims read by deepseek-v4.1-flash under definition version 3 (extract-v3) on 2026-10-06.
Classified by jev-1.13-20260917 (gold-guideline-v0.2) on 2026-10-06.
Cited sources are checked by gpt-6-sol.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review