New episode of The Information Bottleneck 🎙️ We always talk about RL, agents, pre- vs post-training. This time we go a layer down - the infrastructure everything actually runs on Benny Chen (@the_bunny_chen), co-founder of @FireworksAI_HQ , on serving models at scale: why the bottleneck is power, not chips, and why idle capacity worth so much money
Users express enthusiasm for the Fireworks AI co-founder episode on power bottlenecks in model serving by sharing the link and inviting feedback on the infrastructure insights.
No Digg Deeper questions have been answered for this story yet.
Love to hear your feedback! The episode - https://www.the-information-bottleneck.com/infrastructure-for-ai-at-scale-with-benny-chen-fireworks-ai/
New episode of The Information Bottleneck 🎙️ We always talk about RL, agents, pre- vs post-training. This time we go a layer down - the infrastructure everything actually runs on Benny Chen (@the_bunny_chen), co-founder of @FireworksAI_HQ , on serving models at scale: why the bottleneck is power, not chips, and why idle capacity worth so much money
