Found
Taking on Nvidia and betting on transformers with Gavin Uberti from Etched
Aug 20, 2024
Gavin Uberti, the co-founder and CEO of the AI chip startup Etched, dives into the transformative power of specialized chips aimed to rival Nvidia. He shares insights on betting big on transformer models and discusses how these innovations can lead to chips that operate significantly faster. The conversation also touches on the environmental impact of manufacturing, the challenges of fundraising, and his journey from dropping out of Harvard to shaping the future of AI hardware.
39:55
Episode guests
AI Summary
AI Chapters
Episode notes
Podcast summary created with Snipd AI
Quick takeaways
- Etched's specialized AI chips are designed for transformer models, promising significantly faster performance than general-purpose competitors like Nvidia.
- The rise of transformer models, exemplified by ChatGPT, has validated Etched's strategy and increased demand for their technologically advanced chips.
Deep dives
The Importance of Specialized AI Chips
Etched is focused on developing specialized AI chips that are optimized for transformer models, which are proving to be pivotal in modern AI applications. Unlike general-purpose chips like GPUs or TPUs, these specialized chips are designed to achieve higher performance levels by exclusively handling transformer models. This design choice allows Etched's chips to operate significantly faster, processing half a million tokens per second efficiently. The company aims to capitalize on the anticipated continued reliance on transformer models in the AI landscape, especially following their increasing use in various generative AI applications.
Remember Everything You Learn from Podcasts
Save insights instantly, chat with episodes, and build lasting knowledge - all powered by AI.
