When people want to argue that AI is a commodity, they confuse two concepts
1. switching costs
2. fungibility
I do agree that LLMs have low switching costs. But they are not fungible - each model is unique with different strengths and weaknesses.
I also disagree that they can increase switching costs by moving in application stack.
For instance, oil and electricty are commodiites because they are fungible, not because switching cost is high.
Whereas clothing has low switching cost but not fungible
They are fungible in the sense that they can accomplish the same task up to a certain acceptable quality.
If I needed to implement a simple function that is well described, all LLMs (for the most part) are fungible. They will have almost identical implementations (sans variable names) and all be fit for purpose.