![\"Researchers](\"https:\/\/dataconomy.com\/wp-content\/uploads\/2026\/02\/1131743.jpg\" "\\"Researchers")
<\/p>\n<\/p>\n
X\u2019s engineering team published the code powering its \u201cfor you\u201d recommendation algorithm last month, with Elon Musk calling it a transparency victory unmatched by other social media companies. Researchers, however, describe the release as a redacted version lacking meaningful insight into the system\u2019s operations.<\/p>\n
The publication occurred in a context where X stands alone among major social networks in open-sourcing elements of its recommendation algorithm. Elon Musk stated, \u201cWe know the algorithm is dumb and needs massive improvements, but at least you can see us struggle to make it better in real-time and with transparency.\u201d He added, \u201cNo other social media companies do this.\u201d This move followed a similar but earlier release in 2023, yet the current version draws criticism for its limitations.<\/p>\n
\nWe know the algorithm is dumb and needs massive improvements, but at least you can see us struggle to make it better in real-time and with transparency. <\/p>\n
No other social media companies do this. https:\/\/t.co\/UMvBlD1ZpV<\/p>\n
\u2014 Elon Musk (@elonmusk) January 20, 2026<\/p>\n<\/blockquote>\n
John Thickstun, an assistant professor of computer science at Cornell University, characterized the code as a \u201credacted\u201d version of X\u2019s algorithm. He expressed concern in an interview with Engadget: \u201cWhat troubles me about these releases is that they give you a pretense that they\u2019re being transparent for releasing code and the sense that someone might be able to use this release to do some kind of auditing work or oversight work. And the fact is that that\u2019s not really possible at all.\u201d Thickstun\u2019s assessment underscores the gap between the release and actual usability for external analysis.<\/p>\n
User reactions emerged swiftly on X after the code\u2019s publication. Creators shared extensive threads interpreting the code to enhance visibility. One post, viewed more than 350,000 times, advised that X \u201cwill reward people who conversate\u201d and urged users to \u201craise the vibrations of X.\u201d A second post, exceeding 20,000 views, asserted that \u201cposting video is the answer.\u201d Another recommended sticking to a \u201cniche\u201d because \u201ctopic switching hurts your reach.\u201d These interpretations proliferated despite the code\u2019s constraints.<\/p>\n
Thickstun cautioned against deriving strategies from the release. He stated, \u201cThey can\u2019t possibly draw those conclusions from what was released.\u201d The code offers limited operational details, such as filtering out content older than one day, which provides a glimpse into post eligibility but leaves most mechanisms inaccessible.<\/p>\n
Thickstun noted that much of the disclosed information remains \u201cnot actionable\u201d for content creators seeking to influence recommendations. This scarcity of practical insights aligns with the redacted nature, restricting applications beyond basic filtering rules.<\/p>\n
A structural shift distinguishes the current algorithm from the 2023 version. The new system employs a Grok-like large language model to rank posts. Ruggero Lazzaroni, a Ph.D. researcher at the University of Graz, explained the prior approach: \u201cIn the previous version, this was hard coded: you took how many times something was liked, how many times something was shared, how many times something was replied \u2026 and then based on that you calculate a score, and then you rank the post based on the score.\u201d<\/p>\n
Lazzaroni detailed the change: \u201cNow the score is derived not by the real amounts of likes and shares, but by how likely Grok thinks that you would like and share a post.\u201d This transition replaces explicit metrics with model-generated predictions, altering the ranking foundation entirely.<\/p>\n
The reliance on a large language model increases opacity, according to Thickstun. He observed, \u201cSo much more of the decision-making \u2026 is happening within black-box neural networks that they\u2019re training on their data.\u201d He continued, \u201cMore and more of the decision-making power of these algorithms is shifting not just out of public view, but actually really out of view or understanding of even the internal engineers that are working on these systems, because they\u2019re being shifted into these neural networks.\u201d<\/p>\n
The latest release omits details previously available in 2023 regarding interaction weightings. That earlier version specified, for example, that a reply equaled 27 retweets and a reply generating a response from the original author equaled 75 retweets. X redacted these weightings, citing \u201csecurity reasons.\u201d This removal eliminates a key quantitative element from public view.<\/p>\n
Absence of training data details further limits understanding. The code provides no information on the dataset used to train the model. Mohsen Foroughifar, an assistant professor of business technologies at Carnegie Mellon University, emphasized this gap: \u201cOne of the things I would really want to see is, what is the training data that they\u2019re using for this model. If the data that is used for training this model is inherently biased, then the model might actually end up still being biased, regardless of what kind of things that you consider within the model.\u201d<\/p>\n
Foroughifar\u2019s comment highlights potential persistent biases originating from training data, independent of model architecture adjustments. Such disclosure would enable assessments of foundational influences on recommendations.<\/p>\n
Lazzaroni, involved in an EU-funded project simulating real-world social media platforms to test alternative recommendation approaches, views research access as highly valuable. His work replicates platform dynamics to evaluate methods, yet he finds the released code inadequate. He stated, \u201cWe have the code to run the algorithm, but we don\u2019t have the model that you need to run the algorithm.\u201d Without the underlying model, reproduction proves impossible.<\/p>\n
Studying X\u2019s algorithm holds broader relevance. Thickstun noted parallels with emerging technologies: \u201cA lot of these challenges that we\u2019re seeing on social media platforms and the recommendation [systems] appear in a very similar way with these generative systems as well. So you can kind of extrapolate forward the kinds of challenges that we\u2019ve seen with social media platforms to the kind of challenges that we\u2019ll see with interaction with GenAI platforms.\u201d<\/p>\n
Thickstun\u2019s observation connects social media recommendation issues to those in AI chatbots and generative systems, where comparable transparency hurdles arise. Researchers anticipate recurring patterns from social platforms in these domains.<\/p>\n
Lazzaroni, experienced in simulating toxic social media behaviors, critiqued optimization priorities: \u201cAI companies, to maximize profit, optimize the large language models for user engagement and not for telling the truth or caring about the mental health of the users. And this is the same exact problem: they make more profit, but the users get a worse society, or they get worse mental health out of it.\u201d<\/p>\n
This perspective frames engagement-driven design as a shared concern across platforms, linking profitability to societal and individual costs. Lazzaroni\u2019s simulations underscore behaviors amplified by such systems.<\/p>\n
The 2023 release included explicit scoring based on quantifiable interactions like likes, shares, and replies, enabling clearer comprehension of ranking logic. The shift to neural network predictions obscures these processes, as internal model computations evade direct inspection.<\/p>\n
Filtering mechanisms, such as excluding posts over one day old, represent one of few retained specifics. This rule ensures recency in feeds, prioritizing fresh content in user timelines.<\/p>\n
User threads post-release exemplify rapid community analysis, though researchers deem them unsubstantiated. Claims about conversation rewards, video prioritization, or niche adherence stem from partial code views, not comprehensive evidence.<\/p>\n
Musk\u2019s transparency claim positions X distinctly, yet redactions on weights and data preserve core proprietary elements. Security justifications for exclusions maintain competitive edges.<\/p>\n
Foroughifar\u2019s focus on training data reveals a critical audit barrier. Biased datasets propagate effects through model outputs, complicating debiasing without source visibility. Lazzaroni\u2019s EU project context illustrates practical research needs. Simulations demand full algorithmic fidelity, unmet by model-absent code.<\/p>\n
Featured image credit<\/strong><\/p>\n","protected":false},"excerpt":{"rendered":"
X\u2019s engineering team published the code powering its \u201cfor you\u201d recommendation algorithm last month, with Elon Musk calling it a transparency victory unmatched by other social media companies. Researchers, however, describe the release as a redacted version lacking meaningful insight into the system\u2019s operations. The publication occurred in a context where X stands alone among […]<\/p>\n","protected":false},"author":1,"featured_media":44738,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[37],"tags":[],"class_list":{"0":"post-44737","1":"post","2":"type-post","3":"status-publish","4":"format-standard","5":"has-post-thumbnail","7":"category-technologies"},"_links":{"self":[{"href":"https:\/\/agooka.com\/news\/wp-json\/wp\/v2\/posts\/44737","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/agooka.com\/news\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/agooka.com\/news\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/agooka.com\/news\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/agooka.com\/news\/wp-json\/wp\/v2\/comments?post=44737"}],"version-history":[{"count":0,"href":"https:\/\/agooka.com\/news\/wp-json\/wp\/v2\/posts\/44737\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/agooka.com\/news\/wp-json\/wp\/v2\/media\/44738"}],"wp:attachment":[{"href":"https:\/\/agooka.com\/news\/wp-json\/wp\/v2\/media?parent=44737"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/agooka.com\/news\/wp-json\/wp\/v2\/categories?post=44737"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/agooka.com\/news\/wp-json\/wp\/v2\/tags?post=44737"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}