Podcast Atlas
Source note Episode guide Original audio Topics: Technology, Culture

No.232 吴汉坤：AI 时代，演员第一个丢饭碗？

Summary

This 三五环 episode has 刘飞 interviewing actor and AI creator 吴瀚坤 about his route from business school and the 《封神》 training camp into coding, Frederick AI, and AI short-film making. It uses 《粉色悖论》 and 《人口异常》 to show how AI Video Production Workflow changes scripting, storyboarding, image generation, editing, music, and performance control.

The durable contribution is an actor-side boundary for generative video. Wu argues that AI can replace simple, homogeneous, technical, or postproduction-like work, but timing, facial micro-rhythm, embodied emotion, life experience, and artistic expression still need a human subject. The episode also frames AI anxiety as something reduced by direct use, workflow knowledge, and a return to what the creator actually wants to express.

Key Claims

  • 吴瀚坤’s path combines acting, coding, and product work: after pandemic-era career interruption, he learned programming, took CS50, worked in consulting, and built Frederick AI as an early agent product for startup tasks.
  • 《粉色悖论》 began as a short role-farewell demo after 吴瀚坤 played Zhu Yangqi in 《陷入我们的热恋》, then expanded into a fuller AI short-film exercise.
  • AI short-film production lowered the need for a traditional crew but still required hard work across script, storyboard, images, video generation, editing, music, visual detail, and repeated tool learning.
  • Actor AI Performance Boundary / 演员与 AI 表演边界 is clearest in rhythm: Wu says a half-second difference in glance, nod, pause, or comic timing can change a performance, and that text prompts struggle to control this reliably.
  • The near-term film path in the episode is not pure replacement: Wu sees more value in recording real actor performance and using AI for background, scene, and postproduction transformations, similar to a stronger green-screen pipeline.
  • Actor Likeness Authorization / 演员脸部授权 separates commercial use of a face from artistic authorship; a third party controlling an authorized face with prompts or motion may create value without making the result the actor’s own performance.
  • AI-generated performers and AI artists are limited by “human flavor”: intentional errors, broken voice, accent, pauses, and awkwardness can be expressive in human performance but read as model failure when generated blindly.
  • Hands-On AI Anxiety Reduction / 亲手使用 AI 缓解焦虑 appears as the episode’s practical advice for film workers: do not chase every new model or anxiety-selling slogan, but use AI directly enough to learn its limits and decide what remains worth doing.
  • The episode’s broader value claim is that AI shifts people from standard workflows toward self-defined purpose; once tools remove some routine execution, knowing what one wants to make becomes more important.

Key Quotes

“不如做吧” - Wu’s shorthand for starting before certainty.

“电影是遗憾的艺术” - Liu’s phrase for why finished films still carry regret.

“AI像Word” - the episode’s tool analogy: the same software does not equal the same expressive ability.

“使用AI是消解焦虑的方法之一” - Wu’s practical answer to AI anxiety.

Connections

Contradictions

  • No settled contradiction is recorded. The episode reinforces existing wiki distinctions between AI assistance, AI replacement, and human authorship in creative work.
  • The source title uses “吴汉坤”, while the body consistently uses “吴瀚坤”; this ingest normalizes the person page to 吴瀚坤 while preserving the title spelling.
  • Tool names such as Cdance/C-DANCE, AI short-film quality claims, industry-protection expectations, and replacement timing remain source-scoped practitioner observations rather than independently verified industry facts.