In this episode of ChatGPT Curious, we break down OpenAI’s newest release: Sora 2, the text-to-video model now living in its own app, Sora. From how diffusion models work to what the new invite-only platform looks like, we’re covering the tech, the ethics, and the implications. You’ll learn how Sora 2 generates videos, why it’s a big deal (for better and worse), and what it means for the future of AI-generated media.

• What Sora 2 is and how it works
• Diffusion explained (drop of dye in water analogy)
• Denoising and the Willy Wonka comparison
• The difference between training and generation
• SlopTok? Sora app features and invite system
• Opt-out policy and data control issues
• Compute cost and environmental impact
• Deepfakes, realism, and misinformation risks
• Personal takeaways and ethical concerns
• Recent ChatGPT use: Sora prompt + Repairing the wine fridge

• Watch the Sora video that I “made”
• Subscribe to the ChatGPT Curious Newsletter: The Curious Companion
• Submit a Question
• Visit the Website
• Feeling curious AND generous? Click here to support the podcast.

• (00:00:00) - ChatGPT Curious
• (00:00:38) - OpenAI's Sora 2
• (00:03:33) - How Video Generation is Made: Diffusion
• (00:08:05) - SORA2
• (00:09:06) - Chat GPT: Sora 2, Unboxing
• (00:14:22) - Is AI Content Bad For You?
• (00:17:52) - OpenAI Rolls Out 'Sora 2' for Video Making
• (00:18:39) - How I Use Chat GPT this Week
• (00:21:58) - How to Diagnose a Wine Fridge Problem with a Multimeter