What's Your Problem? Teaching AI to Build Stuff in the Physical World
Jan 30, 2025
In this engaging discussion, Edward Mehr, co-founder and CEO of Machina Labs, tackles the challenge of enhancing AI capabilities in the physical manufacturing realm. He shares insights on transforming rigid robots into skilled craftsmen, navigating issues in the aerospace sector, and the limitations of 3D printing. Mehr emphasizes the need for flexibility in production, the collaboration of AI and human expertise, and envisions a future where custom manufacturing becomes accessible to all, unleashing creativity across various industries.
AI Snips
Chapters
Transcript
Episode notes
SpaceX's Rocket Families
- Edward Mehr, having worked at SpaceX, points out that they only had two rocket families in over 20 years due to factory rigidity.
- Changing a rocket's diameter necessitates a whole new factory, a massive investment, as seen with SpaceX's Starship.
The Rigid Factory Problem
- The rigid factory problem describes how factories are designed for specific products, requiring significant changes for new designs.
- Economies of scale aren't a natural law, but a result of this inflexibility, as high initial investments must be spread across many identical products.
Falcon 9's Diameter Constraint
- At SpaceX, the fixed diameter of the Falcon 9 rocket led to design challenges, forcing engineers to cram new features into the existing space.
- This illustrates how rigid factory design constraints innovation and necessitates workarounds.