It’s all about grammars, in the plural. Taken in the sense of the creative imagination, grammars will be the raw material, the components of what surrounds us, whether it be architecture or industrial design. From the point of view of an artificial intelligence, it therefore only needs to make an AI model ingest sector-specific grammars so that, on the basis of its immense knowledge, it can engage in completely new creations.
To make the exercise more concrete, we chose to take the example of a Formula 1 car and explore how a generative AI would go about radically rethinking it, while respecting the regulatory framework of the race. Freed from the weight of the past and the cultural shackles of its historical designers, AI would certainly draw on biomimicry to draw inspiration from the aerodynamics of birds or the structure of their skeleton, the lightness and solidity of which are already the stuff of engineers and architects. Along the way, we discover how a system based solely on statistical prediction is capable of such a degree of unconventional thinking.
To talk about it, Control F invited Christophe Tricot, a regular on the podcast. This former AI researcher is the creator of the La ForgeAI studio. He spends his time exploring the potential of large language models.
Listen to this episode and subscribe to Control F on Apple Podcasts, Spotify, Castbox, Deezer, Google Podcast And Amazon Music.
The team: Frédéric Filloux (presentation), Jules Krot (editing and production), Xavier Yvon (editor-in-chief)
Music and design: Léonard Filloux
Logo: Jérémy Cambour/L’Express
Image credit: Chandan Khanna/AFP
How to listen to a podcast? Follow the leader.