when your AI's context window fills up and you hit compact, it squeezes your whole conversation into a short summary. whatever didn't make it to the summary or it's memory is gone
so i never compact before i save things i can use for later
[my claude does the saving. i'm just telling you what i have it save]
what compact actually does
every AI chat has a context window. that's how much of your conversation it can hold at once. like this:
in Claude Code or Codex, when it gets full, /compact turns the conversation into a summary so you can keep going in the same chat
the summary keeps the big picture. it can lose the small stuff: a decision you made way earlier in the chat, the reason you made it, or something your AI got wrong
you can also type what you want kept right after /compact and it'll keep it
when i compact
i usually compact around 50 to 70% full. not at 95% (which is usually default and automated)
or i skip it: i finish the task and roll over to a brand new chat instead
either way, nothing gets compacted before my logs are in.
what "logs"?
I keep logs of every session i have with my AIs, all it's mistakes, fuckk ups, errors, my decisions, feedback etc.
then i synthesize those logs once a week and teach it things it can do better, make it learn how to work with me, set rules, parameters, guards whatever it may be to improve it's 'default' behavior and personalize it to how i need it to work with me.
the 4 things i save first
everything we've worked on so far
finished decisions go into my vault (the notes folder all my AIs read and write to) and my AI's memory
whatever we're in the middle of goes into a plan doc
all my feedback, every decision, and a log of where claude messed up
that 4th one matters the most (you should keep em too)
why the f*ck ups matter
the logs are how my claude gets better at working with ME
when it messes up and i correct it, that correction goes in the log. then we use the logs to teach claude how to improve it's behavior, how it thinks and what it does for me and with me
over time it gets personalized to me and how i work. Personalized Intelligence is where it's at 🙂 check out me agent-personalizer git repo or install is thru terminal. it'll basically set everything up for you the way I have it
npm i agent-personalizermore detailed the logs, better your AI gets over time!
btw, @suavecito585 was the one who first gave me this idea and basically taught me about 'keeping session logs'. so if you ain't following him, you're definitely doing something wrong with your life! ⭐️
have your agent do it
you don't need to build anything for this. before your next compact, paste this into your chat:
before we compact: save everything we've worked on so far. put every finished decision in [your notes app or memory file]. put whatever we're in the middle of in a plan doc. then write a short log of my feedback, the decisions we made, and anything you got wrong and how i corrected it. tell me when it's all saved
then compact, and point it at the plan doc when you pick back up
watch your context. when it hits 50 to 70%, paste the above prompt, check that everything saved, then compact
you'll notice the difference the first time you don't have to re-explain yourself. and btw... compacting a convo saves token usage because AI doesn't have to keep dragging long string of the convo along.
want this wired into your own setup, so your AI saves its own logs and plan docs before every compact? that's what the AI Workflows for Creators call is for.
Hope this helped!
SEE YA ON THE NEXT ONE ♡
Oh btw... i've been posting a lot of my systems as git repos that you can easily install, so check me out on Github ok byeeee.... 😁
