Member-only story
You’re forty minutes into a debugging session. Claude just found the bug, you’re about to ask for the fix, and instead you get a message telling you to wait three hours.
Look back at the transcript and the bug fix wasn’t what drained you. A tangent about naming conventions did. Three separate re-explanations of the same file structure did. A detour into a feature you were “just curious about” did. None of that was the task. All of it counted against the same clock.
Claude’s usage limits aren’t a flat message counter. They’re a dual-layer system that responds to conversation length, model choice, tool use, and how much thinking a task actually needs. Two people on the identical Pro plan can have wildly different experiences: one hits a wall after fifteen minutes, the other works for hours. Rarely is that the plan’s fault. Almost always, it’s the workflow.
This piece covers how Claude’s usage limits actually work, then four habits that determine how far a given plan stretches: planning conversations before you start typing, using memory and chat search instead of re-explaining yourself, matching model and effort level to the task, and splitting heavy work into separate contexts instead of piling everything into one thread. None of it requires a plan upgrade. It requires…