DigitalFUTURES - Carlos Bañón on AI-Native Architecture with FORMAS

In this DigitalFUTURES presentation, architect and FORMAS co-founder Carlos Bañón walks through the evolution of design technology—from AutoCAD to AI—and demonstrates FORMAS, a next-generation platform that merges human creativity with generative AI. Discover how architects and designers are reimagining the design process using smart nodes, real-time sketching, style creation, and multi-modal workflows. Learn why the interface matters as much as the model, how to maintain agency in AI-assisted design, and why the future of architecture depends on preserving human intuition alongside technological acceleration. 0:00 Intro to FORMAS and framing the talk 0:22 – 2:24 From hand-drawing to AutoCAD: evolution of the design interface (1963–1982) 2:24 – 5:31 Early AI experiments: ControlNet, image-to-3D, Nagami furniture/Miami tree structures 5:31 – 9:12 Teaching AI in academia: São Paulo, Singapore, PA Academy, Monterrey workshops 9:12 – 12:23 Building the first fluid interface, then layering in early GPT assistance 12:23 – 14:54 Speculative work ("impossible villas") vs. Carlos's real built project in Alicante 14:54 – 18:00 Why FORMAS was founded: fixing AI hallucination and ComfyUI's node fragmentation 18:00 – 19:33 The core problem: design as communication, and architects leaving the profession 19:33 – 24:54 Platform fundamentals: shapes, notes, sketching, non-linear design, reference-based generation 24:54 – 28:12 Motion, real-world context insertion, detail generation, asset/catalog extraction 28:12 – 33:06 Company milestones (funding, SuperAI win, university adoption) and Drawing Studio styles 33:06 – 35:57 Agent demo (cost estimation), kids using FORMAS, Milan Design Week recap 35:57 – 37:20 Compute growth trajectory and closing vision for the slide portion 37:20 – 44:46 Live demo: Free Form Studio, style notes from references, deconstructing images into assets, applying materials to a real project 44:46 – 51:00 Precise vs. context generation, selective regeneration, texture transfer, agentic chat consultations (Nina) 51:00 – 1:01:30 Multi-format generation: floor plans, mood boards, Narrative Studio videos, custom style "recipes" 1:01:30 – 1:06:44 Vectorizing to CAD, converting images to 3D models, FORMAS Worlds (360° walkable spaces) 1:06:44 – 1:13:00 Trimble/SketchUp integration, Google Earth context insertion, iterative Photoshop-style editing 1:13:00 – 1:14:49 Closing remarks on FORMAS's philosophy 1:14:49 – 1:45:00 Q&A: speculative vs. practical design, the future of the 3D model, aesthetic homogeneity, individual style, urban-scale design, human intuition vs. AI