Goodbye to Sora

Original Article ↗ Hacker News Discussion ↗

Product concept & user reception

  • Many saw Sora as a flashy tech demo or TikTok-style “AI slop” feed rather than a durable product.
  • Common usage pattern: intense experimentation for 1–2 weeks (often with friends/family), then complete drop-off as novelty wore off.
  • Users reported:
    • Impressive moments but a low “hit rate” (e.g., a few good clips out of dozens or hundreds).
    • Long generation times, high content-violation failure rates, and intrusive watermarks.
    • Videos that felt repetitive in style and “uncanny,” making them hard to watch long term.
  • People mostly generated clips and then posted them on existing platforms (TikTok, Instagram, YouTube), rarely consuming Sora’s own feed.

Economics, compute, and business model

  • Thread consensus: video generation was extremely expensive to run and hard to monetize.
  • Standalone AI-video social network had:
    • Weak network effects versus entrenched platforms.
    • Poor consumer willingness to pay versus subscription or ad-based coding tools.
  • Several commenters think compute and cash are being reallocated to more profitable enterprise and coding products.

IP, safety, and legal constraints

  • Heavy guardrails (copyright, nudity, violence) made Sora “no fun” for power users and blocked most viral IP-based content.
  • Disney’s reported exit from a planned $1B Sora-related deal is seen as highly significant; exact contractual details remain unclear.
  • Some argue IP and deepfake risks make freeform commercial videogen almost impossible at scale.

Strategic shift & competition

  • Multiple references to reports that OpenAI is exiting video in products entirely and pivoting around coding and business users.
  • Many see this as:
    • A pragmatic focus on where enterprises actually pay (coding assistants, B2B tools).
    • Or a sign of strategic drift and financial stress.
  • Chinese models (Seedance/Kling, open-weight Wan/Hunyuan) and Google’s Veo are widely cited as more capable or less restricted; ByteDance is expected to fill the gap rapidly.

Broader implications & ethics

  • Some view Sora’s end as an early “AI bubble” deflation and proof that AI video social apps lack real demand.
  • Others remain bullish on AI video for advertising, stock footage, and pro workflows, but not as a consumer feed.
  • Ethical worries dominate: deepfakes, revenge porn, propaganda, erosion of trust in video, and growing “AI slop” saturating culture.