I almost never run out of tokens

Hi everyone, I have the Claude Pro subscription and I rarely run out of tokens. It feels like I’m doing not doing enough, but I do work the entire day (with some breaks inbetween). Here’s what I do:

Spend most of my time planning the workflow, creating the PRD and iterating over it by chatting with Claude (usually in layer 1, to not spend unnecessary tokens on Claude code). When I’m satisfied, I have it generate markdown files for Claude code.
In Claude code I let fable 5 read the markdown files. After that, it has a step by step plan for me on how to create my end product. I use plan mode to plan through each step and I account for verification steps on the way
When executing tasks, I use Sonnet 5. I use fable 5 to find gaps or possibilities for where it could go wrong and proposing solutions. I rarely use Opus anymore, it feels like I’m going in circles when I’m talking to Opus.
I spend most of my time trying to understand what Claude is telling me. I don’t want to be mindless about my project but it gets really exhausting to read and understand everything it’s doing (It already knows I don’t like technical jargon thanks to the personal assistant that I created🤣).
I keep asking for clarification. I spend most of my time learning while doing.

I’m wondering how everyone else experiences it or has experienced it? I’m quite a noob when it comes to Claude code, but I know quite some things about AI fundamentals such as how context works and how it can waste tokens, as well as how different models work and where they shine (cost-benefit ratio).

But I cannot shake the feeling that something must be wrong, because I keep hearing complaints online about people hitting their limits on Max subscriptions whereas I’m fine with Pro😅

31 comments

I almost never run out of tokens