Jev AI, OpenAI teaser & Opus 5.2 reports - reuse AI feedback | LAB 0001
Lenny's AI Builders - LAB 0001
i kept correcting the same AI mistakes. i asked Codex to find the feedback i had repeated, trace it to the source, and turn it into a workflow i could inspect.
- pick one job where the same mistakes keep coming back.
- add the threads that contain the work and your corrections.
- read the proposed rules before Codex changes anything.
- test the smallest useful version on one real task.
- compare the result with your old corrections.
- keep only the rules that stop a repeated mistake.
i tried this on my job-search workflow. Codex proposed five roles: manager, researcher, writer, designer and reviewer. that is one setup for one task, not a rule for every job.
copy this prompt
read these threads: [add links]
find the feedback i repeated. quote each correction and link it to the source. separate lasting rules from one-off requests. tell me what you could not read.
show me the smallest workflow that would stop the same mistakes:
1. the rules, who owns each step, and each handoff.
2. the final check before the result reaches me.
show me the plan before changing anything.the reviewer found that all 10 evidence-map claim units repeated resume content verbatim. zero were unique to the artifact.
that receipt caught repetition before the application reached me.
the application was not finished or submitted, and i did not measure time saved.
keep it small
more agents means more tokens and more handoffs. start with better rules and one final check. add a role only when its handoff has a clear job. if Codex cannot read a thread, make it say so.
the next move
pick one repeated correction and run the prompt. you are looking for one useful rule you can reuse on the next draft.
sources from this episode
Persona at 8:00. the product page shown in LAB 0001.
Jev at 5:22. the speed and cost figures are the maker's claims.
OpenAI at 0:00. the edit shows the date correction in red.
