This isn’t really about PAYG discipline or being surprised by a big bill. I had a spend cap. The problem is that, because of how Cursor implements caching, that prepaid usage got consumed in a way that was impossible to see or reason about ahead of time. I didn’t “overspend,” I just got far less actual work out of the same budget than I reasonably should have.
Also, the 200k context cap is a real cap. Claude does not attend to more than that. Cursor even summarizes aggressively to stay within it. The confusing part is that billing is driven by something else entirely: replaying cached prompt state that the UI implies is no longer relevant.
So from a user’s perspective, you’re operating inside a 200k window, MAX is off, prompts are small, and things look normal. Yet each call can still replay tens of millions of cached tokens that you never see and can’t inspect without exporting CSVs after the fact.
That’s very different from “you used more than you thought.” It’s more like the primary cost driver is invisible and contradicts the mental model the product presents. If the UI showed cache size or warned that a call would replay ~20M cached tokens, I’d agree this is just normal PAYG behavior. But it doesn’t.
If you read my post carefully you can see that the problem is not the cost of the models or the way I use them but something that is not just hidden from the Cursor UI but misleading. It keeps displaying a 200k context limit and also uses summaries to manage the context. You can only see that it pulled 21M tokens worth of data from Anthropic cache for a 4000 token prompt if you export your usage as a csv and analyse.
I’m doing all of the development and he’s handling product and business. It’s just the two of us. Now that we’re entering an accelerator, it’s basically all or nothing for both of us.
Maybe the better question is this: is it ethically acceptable to expect him to take 0 salary until seed without additional compensation, given that this is the only way we can realistically reach a seed round? In my mindset if he doesn't do this sacrifice he already loses 100% of everything.
Our arrangement over the past year was implied, so yes, effectively a handshake agreement. I don’t have an issue with him taking a salary when we can afford it. The problem is that for the company to maintain runway and keep me full-time, he would likely need to forgo a salary that he doesn’t actually need.
What confused me was how he framed it. First it was about “recouping some of his investment,” and when I pushed back that this isn’t how early-stage compensation works, it shifted to “saving money for family and legacy.” Both are understandable personal goals, but taking salary for those reasons directly increases the company’s risk at a stage where survival comes first. That mismatch is what I’m trying to sort out.
This isn’t really about PAYG discipline or being surprised by a big bill. I had a spend cap. The problem is that, because of how Cursor implements caching, that prepaid usage got consumed in a way that was impossible to see or reason about ahead of time. I didn’t “overspend,” I just got far less actual work out of the same budget than I reasonably should have.
Also, the 200k context cap is a real cap. Claude does not attend to more than that. Cursor even summarizes aggressively to stay within it. The confusing part is that billing is driven by something else entirely: replaying cached prompt state that the UI implies is no longer relevant.
So from a user’s perspective, you’re operating inside a 200k window, MAX is off, prompts are small, and things look normal. Yet each call can still replay tens of millions of cached tokens that you never see and can’t inspect without exporting CSVs after the fact.
That’s very different from “you used more than you thought.” It’s more like the primary cost driver is invisible and contradicts the mental model the product presents. If the UI showed cache size or warned that a call would replay ~20M cached tokens, I’d agree this is just normal PAYG behavior. But it doesn’t.
That’s the issue I’m pointing out.