Podcast Atlas

167: 洋葱学园杨临风：用AI制造捷径，是在杀死真学习

source Updated 2026-07-09 Tags: Podcast, Education, Ai, Learning, Edtech

Summary

This LateTalk episode interviews Yang Lingfeng / 杨凌峰, co-founder and CEO of Yangcong Xueyuan / 洋葱学园, about building a K12 learning company around Self-Directed Learning rather than answer delivery or exam-shortcut training. Yang argues that AI can help as a planner, strategy distributor, and contextual support layer, but only if Learning Experience Design returns students to real reasoning instead of turning AI As Tutor into an answer machine. The source extends the wiki’s AI-and-learning branch by adding AI Shortcut Risk and Education Technology Fairness as constraints on education technology.

Key Claims

  • Yangcong Xueyuan / 洋葱学园 was founded in 2013 after earlier rural education pilots showed that education gaps were not only hardware gaps; learning interaction, content design, and student confidence also mattered.
  • Yang Lingfeng / 杨凌峰 breaks Self-Directed Learning into willingness, ability, tools, and belief: students need to believe they can independently learn, not only receive better explanations.
  • The episode argues that many K12 subjects are learnable for ordinary students, but system-two thinking is effortful and easily avoided when the experience produces repeated failure.
  • Learning Experience Design is Yangcong’s answer to that difficulty: 5-to-8-minute lessons, animation for abstract concepts, purpose cues, empathy for stuck students, achievement loops, and data-driven iteration are all used to lower the threshold for real thinking.
  • Yangcong avoided photo-solution search, two-teacher classes, and large livestream courses because those forms either over-serve students who can pay or move offline classroom logic onto a screen without creating a digital-native learning experience.
  • Current large models are useful as contextual support, but Yang says they still struggle to teach system-two K12 knowledge from scratch with enough child-cognitive adaptation.
  • AI in education has three broad uses in the source: standardizing teacher workflows, automating learning by-products such as homework correction, and making the core learning experience more flexible and vivid.
  • The central risk is AI Shortcut Risk: if AI only follows students’ system-one preference for less effort and faster answers, it can manufacture shortcuts that weaken true learning.
  • Schools and teachers remain important because many students need an external environment to build self-learning habits, confidence, planning, and responsibility.
  • Yangcong’s public-interest structure is not an after-profit add-on: rural schools and teachers meeting the company’s standard have had free access from the beginning, and the source frames mission, product, and business structure as a Day One alignment problem.
  • Yang says the future education goal is not only high scores, but self-reliance: confidence, learning ability, self-discipline, planning, goal-setting, and the ability to learn and apply new things.
  • Knowledge remains important in the AI era because it trains ways of thinking; mathematics is singled out as a carrier for logic and understanding complex systems.

Key Quotes

“创设学习体验，激发每个学生驾驭自主学习” — Yangcong’s mission as stated by Yang.

“每个学渣心中都有一颗想成为学霸的心” — Yang’s counter-consensus claim about ordinary students’ desire to learn.

“用AI制造捷径，是在杀死真学习” — the episode’s title-level warning about AI and learning.

Connections

Contradictions

  • No direct contradiction found.
  • Source-local naming note: the file metadata/title uses “杨临风”, while the transcript body consistently uses “杨凌峰”; this wiki uses Yang Lingfeng / 杨凌峰 for the entity page.
  • Productive tension to track: AI As Tutor and Water And Fire Education emphasize personalized exploration, while this source warns that AI tutoring becomes harmful when it removes the cognitive struggle that builds understanding.