智力贬值的春节见闻录，与那场正在酝酿的优贷危机

Summary

This Keji Luandun episode uses Spring Festival observations, AI video, movies, copyright, digital celebrities, AI coding, and flower-shop/product experiments to argue that AI may reprice human ability more broadly than earlier “AI content” debates suggested. Its central frame is Intelligence Devaluation: education, coding, professional knowledge, and other middle-class bargaining assets may lose scarcity if AI turns them into cheap executable capability. The episode then extends that labor-market risk into Prime Borrower Credit Risk, asking what happens if the white-collar borrowers treated as high-quality credit become income-unstable because the jobs that justified their credit models are repriced.

Key Claims

• Spring Festival robot performances and Seedance-style video generation are treated as capability shocks: robotics and AI video are no longer only lab demos or crude samples.
• The hosts argue that Video Models can move from storyboards and samples toward direct content production, pressuring advertising, short drama, and film-cost structures.
• AI copyright disputes will likely move from simple prohibition toward authorization, labeling, and revenue-sharing, but generated likenesses and famous IP remain legally and emotionally sensitive.
• Digital celebrities can exist as IP, but celebrity value is not only face or acting skill; it also depends on works, roles, memory, and long-running audience interaction.
• GLM5-assisted coding makes implementation feel cheap, but App Store setup, deployment, registration, pricing, audit, operations, and platform know-how remain hard.
• The episode’s “2028 crisis” discussion frames AI as a threat to knowledge-worker bargaining power, not just to one profession or one software task.
• Intelligence Devaluation could flatten output differences between people with different educations or skills, creating an “AI equality” effect while also reducing the market premium once attached to scarce expertise.
• Prime Borrower Credit Risk follows from that labor thesis: if professional income becomes less stable, banks and consumer-credit systems may need to reconsider who counts as a low-risk borrower.
• The hosts worry that model-generated tools and content could become boring or homogeneous when many products are assembled from the same capability layer.
• The short-term refuge is not generic tool operation but vertical Domain Expert Alignment: industry know-how, field observation, user interviews, and knowing what customers actually do.
• The podcast-editing and flower-shop examples show that AI products become valuable when built around real workflow language, pain points, and edge cases, not only around raw model capability.
• Human value shifts toward observation, communication, social trust, demand interpretation, and translating fuzzy problems into executable plans.
• AI-era companies may keep humans at the front end for sales, relationship work, requirement discovery, and exception handling while automating much of the production back end.
• The capital-market discussion connects AI repricing to infrastructure and public equities: AI-native companies may gain attention while electricity, compute, and power equipment become bottleneck themes.

Key Quotes

“智力贬值” — title-level frame for AI making formerly scarce knowledge work cheaper.

“优贷危机” — credit-risk worry that high-quality borrowers may be less stable than old models assume.

“AI 平权” — host shorthand for AI narrowing output gaps between people with different formal credentials or skills.

“中国最好的播客剪辑工具” — internal-test feedback used to show why domain-specific workflow knowledge matters.

Connections

• Keji Luandun — show context for the episode’s AI, media, product, labor, and market discussion.
• Intelligence Devaluation — central concept: AI reduces the scarcity premium of education, coding, professional knowledge, and white-collar cognitive labor.
• Prime Borrower Credit Risk — “优贷危机” extension from income repricing to credit-model fragility.
• Human Resource Deflation Compute Infrastructure Inflation — broader labor-deflation and infrastructure-demand thesis that this episode makes more personal and credit-oriented.
• Middle-Class Consumption Pressure, Financial Career Risk, and Consumer Loan Risk — existing household-finance and career-risk cluster affected if professional income becomes less reliable.
• AI Programming Engine Shift, AI Engineering Thinking, and AI Coding Verification — coding gets easier, but product, deployment, and verification knowledge remain bottlenecks.
• Video Models, Seedance, and AI Content Provenance — AI video improvement, rights disputes, likeness risk, and disclosure issues.
• AI Content Devaluation — audience attention may drop when content feels mass-produced by AI.
• Offline AI Implementation, AI Visual Merchandising, and Operational Data Capture — flower-shop and fieldwork examples that later become a fuller source branch.
• Domain Expert Alignment, Human Judgment Under AI, and AI Communication Ability — human value shifts toward situated judgment, user understanding, and communication.
• Product Led Willingness To Pay and Customer Pull — AI output becomes valuable only when it solves a real paid problem.
• Pinduoduo, ByteDance, Zhipu AI, GLM5, DeepSeek, Baidu, Xiaomi, and Mickey Mouse — product, company, model, investment, and IP examples mentioned in the episode.

Contradictions

• No direct contradiction identified. The source reinforces earlier wiki claims that AI lowers production cost while increasing the importance of domain knowledge, validation, and human judgment.
• It adds a sharper tension to Middle-Class Consumption Pressure and Consumer Loan Risk: earlier pages mostly describe pay cuts, consumption discipline, and borrower-side debt behavior, while this source asks whether the lender-side definition of a “quality borrower” could also break when white-collar work is repriced.