
The FAIK Files Gettin' Sloppy Wit It
51 snips
Oct 17, 2025 This week dives into OpenAI's Sora 2, wrestling with copyright issues and user complaints about censorship. There's a shocking Discord data breach involving 1.5 TB of sensitive information. The discussion also highlights a baffling case where a man's access to nuclear secrets was compromised by AI-generated content stored on a government computer. Plus, the hosts explore the tricky world of watermark removers and AI detection methods, leaving listeners to ponder the chaotic blend of technology and ethics.
AI Snips
Chapters
Transcript
Episode notes
Sora's Copyright Gamble
- OpenAI launched Sora 2 with permissive copyright rules and then faced rapid backlash from rights holders and governments.
- Sam Altman promised more granular control and revenue-sharing but remained deliberately vague on timing and mechanics.
Japan Pushes For Opt-In Model
- Japan urged OpenAI to switch to an opt-in model and to compensate rights holders when characters are used.
- The government framed anime and manga as national cultural treasures needing protection from careless reuse.
Scaling Despite Mounting Costs
- OpenAI continues to scale compute capacity aggressively despite costs, signing large power and chip deals.
- Sora and ChatGPT's growth drives massive infrastructure commitments like multi-10GW arrangements.
