Nah, it's just academic slop of the type every journalist has a crack-level addiction to. OpenAI's incentives are the exact opposite: users pay them a flat fee (or nothing) but OpenAI's cost scale per interaction. OpenAI make more money when people subscribe but don't talk to ChatGPT much, i.e. their incentives are the inverse of what Vasan is claiming here.

Ironically, the Stanford psychiatrist is hallucinating some statistically likely words whilst misinforming readers, perhaps in a way that will make them paranoid. It's turtles all the way down.