Apple's Starflow represents a huge change in the AI landscape, as the tech giant unexpectedly open-sourced a revolutionary auto-regressive flow model that generates images and videos up to 15 times faster than traditional diffusion models, while dramatically reducing computational costs. This game-changing move not only gives everyone access to cutting-edge AI technology but also challenges the dominance of closed-source western competitors like OpenAI, Google and Midjourney, offering developers and creators a powerful, free alternative that could finally make high-quality AI generation accessible to everyone without expensive cloud subscriptions or specialized hardware. 🔗 Relevant Links Starflow GitHub - Starflow Examples - #overview ❤️ More about us Radically better observability stack: Written tutorials: Example projects: 📱 Socials Twitter: Instagram: TikTok: @betterstack LinkedIn: 📌 Chapters: 00:00 - Why did Apple create Starflow 00:33 - Scalable Transformer Autoregressive Flow 01:16 - Why this model is different from the others 02:04 - Normalizing flow model explained 02:52 - Why Starflow is so fast 03:30 - Setting up Starflow on Digital Ocean 03:50 - Image generation examples 05:04 - Going through video examples 05:21 - Can Starflow run on an iPhone? 03:18 - Final Thoughts & Conclusion