History of Generative Media with Fal.ai

Today we are joined by Gorkem and Batuhan from Fal.ai, the fastest growing generative media inference provider. They recently raised a $125M Series C and crossed $100M ARR. We covered how they pivoted from dbt pipelines to diffusion models inference, what were the models that really changed the trajectory of image generation, and the future of AI videos. Enjoy!

00:00 - Introductions
04:29 - History of Major AI Models and Their Impact on Fal.ai
07:06 - Pivoting to Specializing in Diffusion
10:46 - Writing CUDA Kernels
15:50 - Latency Importance and A/B Testing Results with Customers
17:56 - Influence of Open Model Availability on Fal's Growth
19:00 - Working with Closed Source Model Providers
21:19 - Inference Optimization for Audio and Music Workloads
29:10 - Performance Improvements for Video Generation
29:47 - OpenAI and Gemini's Autoregressive Image Generation
34:45 - World Models for Controllable Video Generation
36:26 - Rise of Chinese Open-Source Video Models
39:30 - Monetization Strategies & Revenue Sharing
42:48 - NSFW Content Moderation and Enterprise Content Safety
45:10 - Trends in Startup Launch Videos and Generative Video Adoption
46:59 - LoRA-Based Customizations
47:11 - ComfyUI, Chaining Models, and Enterprise Workflows
51:58 - Applications of Generative Media
54:15 - Requests for Startups and Future Opportunities
56:34 - Ideas for Building Startups on Top of Fal
1:00:29 - Hiring and Team Building at Fal.ai
1:03:27 - What Makes a Cracked Engineer