All articles tagged with #algorithm
Meta's Threads is growing, but cross-posting from Instagram and Facebook can cause unintended posts to appear on Threads, sometimes going viral. Meta has added friction to curb accidental cross-posting, yet the platform's highly personalized algorithm can amplify these misposts, illustrating a divide between users who understand the app and those who don’t.
The article traces algebra to al-Khwarizmi’s 9th‑century Baghdad treatise on calculation by completion and balancing, and shows how the word algorithm comes from the Latinized form of his name in a separate Hindu-Arabic arithmetic text; it also explains how al-Khwarizmi blended Greek geometry with Indian numerals at the House of Wisdom, how his work spread to Europe via translators, and how the terms for both the field and computational procedure evolved into their modern meanings.
Instagram expanded its algorithm personalization to the main feed, letting users select topics they want more or less of; however requests to see more posts from people they follow currently error, underscoring a focus on recommendations, while Meta's Adam Mosseri says the feature—enabled by large language models—gives users more agency and hints at deeper or bespoke personalization in the future.
Instagram is launching 'Your Algorithm' for the main feed, letting users see the topics the app thinks they’re interested in and tweak them—potentially including moods and content types—as Adam Mosseri says the goal is to give people real control over what they see and lay the groundwork for broader, AI-assisted customization.
Instagram unveiled new ranking rules that prioritize original content across photos, carousels, and Reels, making aggregator or repost-focused accounts less likely to appear in recommendations. ‘Original content’ means materials you wholly created or that reflect a unique perspective, including heavily edited third‑party content. To reverse an ineligible status, most of an account’s posts over 30 days must be original, with options to remove unoriginal content or appeal; creators who frequently reshare can stay eligible by sharing others’ posts to Stories, using the repost tool, or enabling collab posts.
Instagram announced a major algorithm update to boost original creators by de-emphasizing aggregator and repost content in Explore and other surfaces, using a rolling 30-day originality assessment. Original content can include edits, photo series, how-to guides, and other value-added posts; third-party material is allowed if the poster adds new value. Reposts outside built-in sharing tools may see reduced reach. The change expands a 2024 Reels initiative, does not affect how content is shown to followers or licensing publishers, and encourages creators to use collabs, remix, or branded content labels to credit others.
Want a calmer X feed? Skip the For You page and stay in Following, toggle between popular and recent, and use Lists to curate who you see. Refine your interests in Settings > Privacy and Safety > Content You See, and train the algorithm by marking Not interested in posts. Muting accounts or keywords helps suppress unwanted content, and being mindful of what you click nudges the feed away from engagement bait. If needed, you can wipe your history with tools like Redact or Deletetweets. In short: tailor your settings and behavior to build a feed that matches your preferences rather than the platform’s default mix.
Axios outlines five practical ways to clean up YouTube recommendations: remove unwanted videos or channels, pause or delete your watch history, actively watch content you want to see more of, log out to browse as a guest, or delete all history to reset and recalibrate your feed.
A 7-week field experiment with US X users shows that switching from a chronological to an algorithmic feed increases engagement and shifts political attitudes toward conservative positions on policy priorities, Trump investigations, and the Ukraine war; switching back to chronological has no reversing effects. The algorithm also promotes conservative content and activist accounts while demoting traditional media, and people tend to continue following these accounts after the switch, explaining why prior studies found no political effects when turning the algorithm off. Partisanship and affective polarization were largely unaffected. Compliance was high, and effects were stronger among Republicans and Independents.
The Verge profiles Nick Shirley, a 23‑year‑old influencer whose viral fraud video about daycare fraud helped trigger federal action, using it to illustrate how cheap, sensational, algorithm‑driven “slop” content functions like modern yellow journalism and shapes politics today. The piece argues that independent creators can wield outsized influence, raises concerns about journalism ethics in the age of data‑driven performance, and highlights the uneasy line between being seen as “journalists” and being part of the political power ecosystem.
Ars Technica reports that TikTok users’ fears of censorship are amplified by what appear to be technical issues (upload failures and DM problems) that coincide with anti-Trump, Epstein, and anti-ICE content removals. Experts say even if glitches exist, the consistent pattern suggests biased moderation, fueling concerns about political content suppression and trust in the platform. Some users are deleting or pausing use, while others test boundaries and seek workarounds; TikTok attributes issues to a US data‑center outage, is investigating, and has updated terms of use and privacy policy. Long-term, the situation could erode trust and reshape TikTok’s role in political discourse as the new US ownership oversees algorithm tweaks and content moderation policy.
TikTok’s US outage coincides with the company’s transfer to majority-US investors, fueling fears that the new owners could influence the algorithm or suppress political content; TikTok says the disruption is a power outage at a US data center, with no evidence of censorship, while engagement and posting were affected. The episode heightens scrutiny of the new USDS Joint Venture and Oracle’s stake, though details on governance and data practices remain unclear.
A deal to install a majority-American board for TikTok’s U.S. operations puts Trump-friendly investors in charge of a major cultural platform used by about 170 million Americans, sparking debate over whether ownership could steer the app’s algorithm toward pro‑Trump content. The arrangement, brokered with national security in mind, reflects broader shifts in U.S. media ownership and raises questions about the platform’s influence on younger voters, while observers stress that the exact impact on content remains uncertain.
The US finalizes a TikTok deal creating a US-based joint venture with ByteDance holding 19.9% and investors including Oracle, Silver Lake, and MGX; Americans gain majority board control and US data will be retrained and housed in a US cloud, but critics warn the arrangement may not fully separate ByteDance from TikTok’s algorithm, leaving ongoing security concerns and potential MAGA-driven content decisions under congressional scrutiny.
TikTok announced a US-led Joint Venture (TikTok USDS) to run the app in the United States, with ByteDance retaining a 19.9% stake and Oracle licensed to manage and retrain the recommendation algorithm on US user data in a secure Oracle cloud. Most users won’t need to download a new app, but the For You feed could evolve as the algorithm is updated on US data, and terms now include under-13 limits and AI-use disclosures. The specifics of how user experience will change remain unclear, though the company says it will preserve a “global experience” for US users and extend safeguards to CapCut and Lemon8.