Owning the Inference Stack 01 with Vidya - Open Registration

​Part 1: The Silicon

​Owning the Inference Stack is a technical series on how modern AI actually runs, from the metal up. Three parts:

• ​Part 1: The Silicon (tonight)

• ​Part 2: The Architecture

• ​Part 3: How to Train Your drAIgon

​Tonight is Part 1. A proper technical night on how the silicon that powers inference actually works. By the end you should be comfortable asking the question that decides every workload: what boundary is it running into, compute, memory, bandwidth, network, or software? We get into the physics of silicon computing to better understand the real constraints of AI, and look at the hardware shaping the field right now: NVIDIA, Cerebras, Groq, TPU, Trainium.

​Free and in person at Frontier Tower, 9th floor Annex, 995 Market Street.