Podcast Atlas
entity Updated 2026-08-16 Topics: Technology

Seedance

Seedance is the ByteDance video-model product referenced in the wiki’s AI-video sources. Vol. 162 科技快乐星球44: 新模型“SOTA们”齐贺新春 treats Seedance 2.0 as a strong Video Models signal because of clearer images, more cinematic movement, stronger camera behavior, and visible international demand.

「蜘蛛侠」新片拿下近半国内票房，AI 模型爆发价格战 adds C-DANCE 2.5 as a current product-update point in the same ByteDance video branch. The source says single-generation duration doubled to 30 seconds and that the model improved long narrative ability, multimodal reference handling, and editing; it also says C-DANCE had been used at scale in China’s short-drama industry and had taken some overseas market left open after Sora.

智力贬值的春节见闻录，与那场正在酝酿的优贷危机 adds an earlier practical interpretation: hosts discuss ByteDance video generation as moving from mockups and sample production toward content that can replace parts of filming, advertising, and customer-facing presentation.

140. 对姚顺宇的4小时访谈：请允许我小疯一下！在Anthropic和Gemini训模型、技术预测、英雄主义已过去 adds Yao Shunyu / 姚顺宇’s cautious interpretation: Seedance can pressure peers through product effect and data/execution details, but he does not read it as a clear paradigm shift in Video Models.

266.从红果到AI短剧：谁在革谁的命？ refers to “C-dance” in a short-drama production and data-flywheel context. The wiki records this as a Seedance-style extension rather than assuming every naming detail: the episode’s relevant claim is that large numbers of creators repeatedly generating AI short-drama material can feed back into Video Models improvement.

267.3000块成本，3.5亿次播放，AI短剧怎么在抖音挣钱？ adds a creator-side production example. 小果哥哥 / XiaoGuoGege says Seedance 2.0 and adjacent video tools helped make 安徽小木匠 / Anhui Xiao Mujiang feasible for a solo creator after earlier tools such as Sora felt too weak on character consistency.

从央视纪录片到爆款 AI 短剧：第一批「转身」的导演 | S10E11 adds a director-side usage case. 抽象仔 / Chouxiangzai treats C-DANCE 2.0 as the point where clients and creators began to see stronger AI-video realism, better multi-reference handling, and unexpected details that can reduce the need for low-level “draw-card” labor inside AI Director-Core Workflow.

175: 对话Liblib陈冕：关于活下来，以及所有接近死亡的时刻 adds a downstream pricing controversy through Lib TV. Chen Mian / 陈冕 denies that Lib TV’s low price should be read as a simple discount to the Seedance API price, arguing that subscription credit consumption, renewal, and LTV determine the product economics.

Key Points

  • Seedance is used as evidence that AI video is crossing from prototype output toward production-like media.
  • The model is discussed through examples such as flower-display videos, AI short dramas, advertisements, and film-like shots.
  • Better video generation increases creative leverage but also intensifies AI Content Provenance, copyright, likeness, and IP-licensing questions.
  • In the episode’s broader labor thesis, video generation contributes to Intelligence Devaluation by making once-specialized production skills cheaper.
  • Episode 140 treats Seedance as impressive product execution without equating it to a new model-training paradigm.
  • Episode 266 adds the AI short-drama usage loop: creator “draws,” platform feedback, and production demand can become model-improvement signals.
  • Episode 267 adds the low-budget AI short-drama case where improved video generation made a 19-day production cycle plausible, while editing still remained the bottleneck.
  • The What’s Next source adds that C-DANCE 2.0-style multi-reference generation can shift labor away from raw image selection and toward director intent, assets, storyboards, and performance judgment.
  • The LateTalk source adds Seedance as an upstream API reference point in Lib TV pricing debates, not as the full explanation for the application business.
  • The 声动早咖啡 source adds C-DANCE 2.5’s longer generation duration and commercialized short-drama usage as a fresh video-model competition signal.

Connections