If all you need to do is deploy AI without training it or developing the algorithm/architecture, then yes, fast.ai is fine. The problem occurs when people confl

If all you need to do is deploy AI without training it or developing the algorithm/architecture, then yes, fast.ai is fine. The problem occurs when people conflate the knowledge needed for deployment with the knowledge needed for development.