[scuderiaseb]

I must be doing something very different from everyone else, but I write what I want and how I want it and Opus 4.7 plans it for me, then I carefully review. Often times I need to validate and check things, sometimes I’ve revised the plan multiple times. Then implementation which I still use Opus for because I get a warning that my current model holds the cache so Sonnet shouldn’t implement. And honestly, I’m mostly within my Pro subscription, granted I also have ChatGPT Plus but I’ve mostly only used that as the chat/quick reference model. But yeah takes some time to read and understand everything, a lot of the time I make manual edits too.

[wg0]

>Then implementation which I still use Opus for because I get a warning that my current model holds the cache so Sonnet shouldn’t implement.

This is based on the premise that given detailed plan, the model will exactly produce the same thing because the model is deterministic in nature which is NOT the case. These models are NOT deterministic no matter how detailed plan you feed it in. If you doubt, give the model same plan twice and see something different churned out each time.

> And honestly, I’m mostly within my Pro subscription, granted I also have ChatGPT Plus but I’ve mostly only used that as the chat/quick reference model. But yeah takes some time to read and understand everything, a lot of the time I make manual edits too.

I do not know how you can do it on a Pro plan with Claude Opus 4.7 which is 7.5x more in terms of limit consumption and any small to medium size codebase would easily consume your limits in just the planning phase up to 50% in a single prompt on a Pro plan (the $20/month one that they are planning to eliminate)

[scuderiaseb]

> I do not know how you can do it on a Pro plan with Claude Opus 4.7 which is 7.5x more in terms of limit consumption and any small to medium size codebase would easily consume your limits in just the planning phase up to 50% in a single prompt

I also don’t understand because all I ever hear is people saying $100 Max plan is the minimum for serious work. I made 3-4 plans today, I’m familiar with the codebase and pointed the LLM in the direction where it needed to go. I described the functionality I wanted which wasn’t a huge rewrite, it touched like 4 files of which one was just a module of pydantic models. But one plan was 30% of usage and I had this over two sessions because I got a reset. I did read and understand everything line of code so that part takes me some time to do.

[aforwardslash]

One of the simple "reasons" is to keep context clean; if you're doing planning, you're not loading source code, its just the plan. Also, it may happen that if you're running parallel manual sessions, cache expires after 1h, so a prompt on an idle session will re-trigger re-evaluating the whole context (something quite heavy on a 1M context window). This burns a lot of credit.

[_puk]

Rather than vibe, write your thoughts and get the model to challenge you / flesh it out is my preferred approach.

Get it to write a context capsule of everything we've discussed.

Chuck that in another model and chat around it, flesh out the missing context from the capsule. Do that a couple of times.

Now I have an artifact I can use to one-shot a hell of a lot of things.

This is amazing for 0-1.

For brown field development, add in a step to verify against the current code base, capture the gotchas and bounds, and again I've got something an agent has a damn good chance of one-shotting.