Oh man… I dunno what to say. Do you get it to discuss before writing code? Or set up a CONCEPT.md file or ask it to confirm before coding & update the CHECKLIST file before editing the code?
https://x.com/vveerrgg/status/1870241281622016328?s=46&t=AzmbXTepQ7mKGb9E0dbu1Q
🤖 Tracking strings detected and removed!
🔗 Clean URL(s):
https://x.com/vveerrgg/status/1870241281622016328
❌ Removed parts:
?s=46&t=AzmbXTepQ7mKGb9E0dbu1Q
Are you messing with me? Because I’m pretty sure this won’t work 😂
I wouldn’t be so reactive and manically posting if I wasn’t so amp’d on it.
2 months ago I was in your exact spot. Then I started talking to the AI LLM saying we were pair programming & I was relying on it to help write good code. Then we agreed we needed a “memory document” for when it would deviate from the plan … and in the end we landed on a CHECKLIST file as our Rosetta Stone.
It just ignores my checklist lol
It’s a mix of those “super prompts” ppl talked about a couple of months ago … and also my response to the LLM API closing the connection if I idle too long or saturate my tokens in a fast moving agent flow
