If I was in charge, I would make the service really cheap and effective, get tons of early adopters, and grow a massive user base all at a loss. Then once the market share is secure, I would try to inch the price up (or slowly remove value in features, which is the same as increasing the price) to capitalize on the promotional “marketing” campaign of gaining a large market share early on. Of course I could not tell the customers this, so I will just make changes to the service gradually instead of all at once, and offer unlimited auto to soothe the price changes and to secure annual membership funding, then remove Unlimited Auto for new customers, and gradually reduce the effectiveness of Auto for anyone grandfathered into it until they are forced to upgrade beyond Pro or pay per token. I’d deploy a few people to the forums to constantly reassure people what is going on, all while knowing in three weeks it will all change again. The customer won’t know they’re getting boiled. Profit!
But in all honesty, I am just going to use Cursor until the cost becomes greater than the value, or a similar competitor appears and is worth the cost of changing workflows.