Scarcity of tokens and pay to play

The cursor pricing has been a bit confusing over time. Recently I circled back to my account usage and their documentation.

It turns out I have free usage of “Auto” until the end of my current billing cycle. They were giving that benefit in return for using the traffic to improve internal model selection.

When I realized that, I became much looser with firing off background agents on various tasks. When I get back to the regular plan we’ll see if I’m going to brush up against limits.

If companies stop subsidizing tokens, we’ll see more of a scarcity model take hold for a while. I am a little more careful and thoughtful when I’m working within the monthly quota.

Of course others will just pay their way through the tokens or expense then costs.

At some point, optimization and economies of scale may reduce the scarcity concern. Until then, there’s a little bit of a pay to play effect here on leveraging AI.