Training Data How Ricursive Intelligence’s Founders are Using AI to Shape The Future of Chip Design
102 snips
Jan 14, 2026 
Azalia Mirhoseini and Anna Goldie, co-founders of Recursive Intelligence, revolutionized chip design at Google with AlphaChip, drastically speeding up the process. They discuss how chip design bottlenecks hinder AI's progress and their vision for 'designless' custom silicon, making it accessible for all companies. The duo shares insights into using AI for advanced placements, novel chip shapes, and recursive self-improvement, where AI enhances its own designs. Their optimism for AGI and a 'Cambrian explosion' of custom silicon applications paints an exciting future for technology.
AI Snips
Chapters
Transcript
Episode notes
Winning Trust With Weekly Demos
- The TPU team required weekly demos for years to build trust and adoption of AlphaChip layouts.
- Gradually AlphaChip was used on more area and achieved increasingly superhuman performance across generations.
Performance Scales With Data And Recursion
- AlphaChip's performance delta grew with more training data, showing scaling benefits of data-driven design.
- Recursive Intelligence names itself to capture the loop of AI designing chips that then accelerate AI.
Chips Enable Recursive AI Progress
- Faster, more custom chips bend AI scaling laws and accelerate the development of stronger models.
- This creates a recursive self-improvement loop: better chips fuel better AI which produces better chips.