How is any of what you wrote relevant? People aren't using Claude for the first time and hitting rate limits. They've been using Claude for months, at the very least, and they're hitting rate limits without significant changes to how they prompt.
> People need to understand a few things: vague questions make the models roam endlessly “exploring” dead ends.
> If people were considerably more willing to aggressively prune their context and scope tasks well, they could get a lot more done with it
If this were the problem, people would've encountered this when they started using Claude. The problem is not that they can't get anything done. It's being able to get things done for months, but suddenly hitting rate limits way too easily and response quality being clearly degraded, so they can't get things done that used to be possible.
> People need to understand a few things: vague questions make the models roam endlessly “exploring” dead ends.
> If people were considerably more willing to aggressively prune their context and scope tasks well, they could get a lot more done with it
If this were the problem, people would've encountered this when they started using Claude. The problem is not that they can't get anything done. It's being able to get things done for months, but suddenly hitting rate limits way too easily and response quality being clearly degraded, so they can't get things done that used to be possible.