# World’s Fastest AI Inference: A Conversation with SambaNova’s Innovators

Page: https://stenobird.com/podcast/generative-ai-meetup/world-s-fastest-ai-inference-a-conversation-with-sambanova-s-innovators
Text version: https://stenobird.com/podcast/generative-ai-meetup/world-s-fastest-ai-inference-a-conversation-with-sambanova-s-innovators.md
Podcast: [The Generative AI Meetup Podcast](https://stenobird.com/podcast/generative-ai-meetup)
Published: 2024-10-25T01:11:11+00:00
Episode link: https://podcast.genaimeetup.com/e/world-s-fastest-ai-inference-a-conversation-with-sambanova-s-innovators/
Audio file: https://mcdn.podbean.com/mf/web/gfenxerq9796pydx/Oct_24_-_Sambanova-enhanced-85pbdtx2.mp3
Processing state: failed
JSON: https://stenobird.com/v1/public/podcasts/generative-ai-meetup/episodes/world-s-fastest-ai-inference-a-conversation-with-sambanova-s-innovators
Duration seconds: 3805

## Resource

This week, Shashank and Mark sit down with SambaNova Systems, a leading AI chip startup competing with tech giants like Nvidia and Cerebras. Joined by SambaNova's Director of Machine Learning, Urmish, and founding team member Raghu, they explore how SambaNova's reconfigurable data flow architecture is changing the game in AI inference and training. They discuss the company’s unique hardware, fast inference capabilities, memory optimizations, and what the future holds for AI chip innovation. Learn what it takes to build high-performance AI systems and where the industry is headed next!

## Actions

- request_transcript: `POST https://stenobird.com/v1/public/podcasts/generative-ai-meetup/episodes/world-s-fastest-ai-inference-a-conversation-with-sambanova-s-innovators/transcription-requests` — Idempotently request low-priority transcript generation for this episode.
- read_markdown: `GET https://stenobird.com/podcast/generative-ai-meetup/world-s-fastest-ai-inference-a-conversation-with-sambanova-s-innovators.md` — Read the agent-friendly Markdown representation of this episode resource.

A page view does not enqueue transcription. Agents should invoke `request_transcript` explicitly when they need this episode processed.

## Transcript

Full transcripts are not published on public pages unless there is a clear rights basis.