Sora, AI Bicycles, and Meta Disruption (stratechery.com)

OpenAI’s new Sora app has quickly become a viral hit, outpacing early impressions and even beating Meta’s Vibes in App Store charts. The surprise isn’t just slick visuals — it’s that Sora makes genuinely shareable, often hilarious short videos super easy to create. Feeds are full of personalized “cameos” and realistic, real-world scenes generated from simple prompts, which turns casual browsers into prolific creators. Where some reviewers (including the author) preferred Meta’s lean-back, curated Vibes experience, Sora’s addictive creative loop has drawn millions who want to make content, not just consume it. That shift matters for AI/ML because Sora exemplifies the next stage of the “AI unbundling”: removing creation as a bottleneck. By turning complex video production into a prompt-driven process, it flips parts of the 90/9/1 dynamic—dramatically expanding who can be a creator and accelerating amateur experimentation (think GarageBand × TikTok). Technical implications include a flood of low-cost, machine-generated content (“AI slop”), new moderation and distribution challenges, and renewed competition among platforms that target creators (YouTube), editors (OpenAI), and passive consumers (Meta). The broader consequence: democratized creativity at scale with both cultural upside (new voices, formats) and trade-offs for craft, quality, and platform economics.