Training vs inference: where AI compute money goes
AI has two bills. Training is the big cheque for teaching a model, and inference is the running bill for every answer it gives. A simple example shows how quickly a penny an answer can catch up with the cost of training.
The example uses made-up numbers. The 90% figure is what customers told AWS in 2019, not a measured industry figure.
More at techarcade.io: https://www.techarcade.io
#AI #Inference #MachineLearning #Startups #shorts


