RoboDojo w/ Rory Flynn - Weavy.ai Masterclass скачать в хорошем качестве

RoboDojo w/ Rory Flynn - Weavy.ai Masterclass

🎯 Episode Overview In this deep-dive RoboDojo session, creative systems designer Rory Flynn breaks down how he builds scalable, repeatable, multi-model AI pipelines using Weavy. From deconstructing images into “non-negotiable” components to constructing modular system prompts and full batch workflows, Rory shows how to turn a single idea into hundreds of high-quality outputs. He also demonstrates compositing, 3D-assisted fixes, character workflows, ad builders, aesthetic testing, storyboarding, and multi-model video prompting. This episode is packed with practical insights for filmmakers, designers, advertisers, and anyone exploring AI-native content creation. 🧩 What You’ll Learn How to think in node-based workflows instead of single prompts Why Weavy is emerging as a creative OS for AI teams Rory’s full method for deconstructing images How to build system prompts that act like internal engines Using batching to scale ideas into dozens or hundreds of outputs The power of the Compositor node for character + scene building How to test one idea across multiple AI models at once When to use 3D tools (Rodin/Trellis) to solve tricky visual problems How to build: Character generators Product ad creators Photo shoot workflows Meme engines Storyboard expanders E-commerce try-on setups Multi-model video prompt lanes 🗂️ Chapter Markers (Timecodes approximate) 0:00 — Welcome & Introductions Palm Springs, architecture, media deadlines, AI agents vs reality 1:50 — Rory: “I’m in four converging projects at once” The holiday sprint, over-booking, and productivity pressures 4:00 — The Agent Bubble Why most “I built an agent in 3 minutes!” posts are lies 5:30 — The Origins of This Session Why Weavy was chosen as the focus 7:00 — Rory’s Background Systematic AI, performance marketing, scale-focused creative tooling 8:25 — The Case for Weavy Why node-based thinking is the next evolution How Weavy differs from Comfy, Flora, and traditional prompting 13:00 — Deconstruction 101 Breaking images into shot/subject/environment/composition/lighting/mood/etc. 17:00 — Turning Building Blocks into Formulas Creating prompt skeletons for consistency and reuse 22:00 — System Prompts as Engines How Rory writes and uses them Role → Goal → Output rules → Format → Limits 30:00 — Workflow Thinking How he reverse engineers from the final output Proofing → system prompt → workflow → scale 33:00 — Inside Weavy Core nodes, UX, model integrations, and why compositing changes everything 44:00 — Branching, Arrays, and 10× Prompting Turning one loose idea into 10–50 prompts automatically 47:00 — Real Workflows (Ads, Characters, Aesthetics, E-com) Full breakdowns with images and composites 1:05:00 — Using Weavy for Film & Storyboarding How to generate consistent scenes and shot lists 1:12:00 — Compositor Deep Dive Layering, scene assembly, character packs, text, props, 3D assets 1:20:00 — Questions from the Community Costs, scaling responsibly, UX tips, replacing mid-journey, model swaps 1:32:00 — Where This All Goes Next Weavy × Figma Agentic workflows Blueprint-based creativity 1:35:00 — Closing & Next Steps 💬 Notable Quotes “Paragraph-and-pray is not a workflow.” “If you crack the system once, the whole team gets the win.” “You don’t need ten subscriptions when one workflow can swap models.” “We’re training these tools how to build agents by using them.” “The compositor node is the cheat code.” 🔗 Tools Mentioned Weavy (primary workflow tool) Flux, Minimaxi, Revid, Imogen, Mystic, GPT-Image Rodin 3D, Trellis Higgsfield / Soul references Runway (for comparison)