Podcast Atlas
source Episode summary Updated 2026-08-06 Tags: Podcast, Algorithms, Media, Recommendation, Attention

164.算法的“兔子洞”：为什么你总在看完新闻后滑向娱乐？｜对谈黄圣淳教授

Summary

This [[QizhulouYanBinke|起朱楼宴宾客]] algorithm-series conversation pairs [[DavidWeng|大卫翁]] with [[HuangShengchun|黄圣淳]] to move from broad Public Relevance Algorithms / 公共相关性的算法 theory into media-effects evidence. The episode argues that algorithmic feeds do not simply trap users in total [[InformationCocoon|information cocoons]]; they also create Incidental Exposure / 偶然暴露, intensify Affective Polarization / 情感极化, route public information toward [[AlgorithmicEntertainmentRedirect|entertainment]], and train users and creators through [[PlatformAffordance|affordances]] and [[PlatformFeedbackLoop|feedback loops]].

Key Claims

  • [[HuangShengchun|黄圣淳]]’s path from news training and recommendation-tagging work to algorithm research frames algorithms as new [[PublicRelevanceAlgorithms|gatekeepers]] between newsroom production and public attention.
  • The episode distinguishes Information Cocoon / 信息茧房, Filter Bubble / 过滤气泡, and echo-chamber worries, then emphasizes that empirical evidence for universal or stable filter bubbles is weaker than public discourse often assumes.
  • Incidental Exposure / 偶然暴露 and News Finds Me / 新闻找到我 complicate cocoon panic: algorithmic or social feeds can make users encounter news and opposing information they did not actively seek.
  • The source still treats algorithmic environments as powerful because platforms preselect available content pools, ranking signals, and interaction forms before users make visible choices.
  • Algorithmic Amplification / 算法放大 can make conflict feel more frequent and more representative than it is, especially when angry comments, disagreement, and discomfort become measurable engagement.
  • Cross-viewpoint exposure may produce Affective Polarization / 情感极化 rather than understanding when users meet opposing views too intensely, too quickly, or in antagonistic platform formats.
  • [[HuangShengchun|黄圣淳]]’s YouTube audit finding is source-scoped but important: even without real user behavior, the recommendation chain more often moved from news to entertainment than from entertainment back to news.
  • The episode treats Algorithmic Entertainment Redirect / 算法娱乐重定向 as a civic problem: public information can lose attention not because it is hidden, but because recommendation systems keep lightly nudging users toward easier, safer, or more engaging content.
  • Algorithmic Cultural Flattening / 算法文化压平 connects platform feedback to culture: fast data, visible reaction, and commercial return pressure can reward safer, more imitable, lower-friction cultural forms.
  • [[PlatformAffordance|Affordances]] such as likes, comments, reposts, autoplay, and infinite scroll are not neutral interface details; they shape the relationship among user, platform, creator, and information.
  • The practical response is Algorithmic Media Literacy / 算法媒介素养: notice when one is giving the system feedback, reduce reflexive engagement with rage bait, compare multiple media environments, and use Feed Curation deliberately.

Key Quotes

“新闻找到我” — the episode’s phrase for news arriving through feeds rather than active seeking.

“轻推” — Huang’s description of entertainment redirection as a subtle shaping force rather than direct coercion.

“意识到自己在做什么” — the user-side literacy principle for dealing with feedback loops.

Connections

Contradictions