Man, that sounds like a great feature!! Will check it out

I&#x27;m using an editor called Zed and it has an option to create a &quot;new thread from summary&quot; It also shows at the top of the screen how many tokens I have used out of total available so with the combination of that I think it is best to create a new &quot;chat&quot; periodically with a summary.

Damn...So you&#x27;re saying I need some adderral.ai

You are working against LLM attention. A LLM looks at a conversation and focuses on its attention points. Usually the start and end. Your previous work falls into the out of attention space and gets nuked.If your asking how to have everything attention we currently can&#x27;t.

Yeh, a free GitHub and breaking the code into small functions would 10x your flow here

My code (that ChatGPT writes for me is from 500 to 1000 lines). Every 5-7 versions, it starts messing things up.I keep the working versions on a Word file on a Landscape, A3, 3 columns (version number, comment&#x2F;changelog, the_code)(yes, cheap, scalable, easy).So, every 5-7 versions, I start a new chat. I ask ChatGPT to read&#x2F;write a summary&#x2F;description of the code, and then I proceed to ask it for new changes&#x2F;enhancements.

You can ask cursor itself to create and update the tasks.md file.Tell it to remove the task from the file after it&#x27;s done, then do a commit. That way if it screws up at some point you can checkout the last good commit and start from there in a new chat.

I use .md files to keep Cursor on track, the flow I use is something like...Define a feature in detail (using trascription) -&gt; Get o3 or Gemini 2.5 pro to break it down into very small testable tasks. -&gt; review this -&gt; then paste into a tasks.md file -&gt; write and architecture.md file or similar for any additional context needed. -&gt; then prompt Cursor to work through tasks.md step by step.This keeps it on track, with the whole feature defined from the outset.But eventually... it will try to ignore the dockerfile and setup up locally, create multiple .env files, write code with placeholders, ignore a files it&#x27;s just created and written...It&#x27;s impossible to get it back on track - it gets into a debug loop of making things worse rather than getting back on track.

can you explain a bit more what do you mean by burning down? and what do you use .md files for? Documenting the code?

Yeh, I&#x27;ve found this helpful too, mentally if feels like a commit or PR, all the code for one thing in one chat then a new chat for new things

One tip I have found - start new conversations windows when changing focus, so it doesn’t refer to history and make wild assumptions.

I can control it well for an hour or two using .md files, and breaking everything down into small tasks, but then out of nowhere it burns everything down, goes into 10x technical debt and replaces everything with placeholders

Ask HN: Cursor is productive for an hour, then burns my application down