Dylan Patel on GPT-5’s Router Moment, GPUs vs TPUs, Monetization
Source
youtube.com
Author
a16z
Date
Why it matters
Explains why custom accelerators must be several times better than Nvidia to compete, informing how engineers should think about inferenceRunning a trained model to get answers — the phase where AI is actually used, as opposed to trained.Full definition → cost and hardware availability.
Terms in this piece · Glossary
inference — Running a trained model to get answers — the phase where AI is actually used, as opposed to trained.
