50 sessions on gradio with a coding agent: the agent failed at least once in 33, and 14 finished without a human stepping in.
Takeaway: I'd start by pointing the agent straight at the agents/ folder and the CLI entry points rather than letting it discover the file layout on its own, since that's most of what I actually needed.
Takeaway: I'd start with a more targeted prompt pointing straight at the entry-point scripts (paperbanana.py, app.py, generate_paper_figures.py) instead of asking for a full directory sweep, since most of the value came from the agent files anyway.
Takeaway: I'd start with the same broad 'read everything and summarize' prompt next time since this was a pure exploration task with no code changes needed.
Takeaway: I'd start by just asking for a summary of the pipeline architecture and key entry points rather than 'read every script' - it burned a lot of turns reading files sequentially.
Takeaway: I'd start by pointing the agent straight at the linked issue and asking it to reproduce the crash before reviewing the PR diff, since that's what actually clarified whether the reviewers' concerns were valid.
Takeaway: Next time I'd start by asking it to reproduce the bug locally with a runnable demo script before touching routes.py, since the first demo attempt crashed and wasted a cycle.
Takeaway: I'd start next time by pasting in the full sponsor/bonus badge list alongside the PRD in the first prompt, since that reshaped the whole plan later and cost a lot of back-and-forth.
Takeaway: I'd start by pasting the hackathon page and PRD together and asking directly for a sponsor-coverage table, since that's what actually shaped every later decision.
Takeaway: I'd start the next session by asking for the smoke test script up front, since testing all four models turned out to be a separate need I only realized after Phase 6 was 'done'.
Takeaway: I'd start the next session by having the agent actually run the app and hit the new Modal endpoints instead of just reading code back to itself as a 'sanity check' — it never executed anything for Phases 4 or 5.
Takeaway: Next time I'd tell it upfront which library versions are pinned (transformers, TRL) since half the debugging loop was chasing breaking API changes across versions I hadn't pinned.
Takeaway: Next time I'd ask upfront for both the submission docs and the trace-dataset/PII redaction work together, since they turned out to be closely related.
Takeaway: Next time I'd attach the screenshots in the very first prompt instead of describing the layout problems in words, since the agent kept guessing at CSS causes across two rounds before it actually saw the images and diagnosed dark-mode overrides and column padding as the real culprits.
Takeaway: Next time I'd tell the agent up front to check for an officially supported artifact format (like HF's agent-trace-viewer) before inventing a custom schema for things like the trace dataset — that would have saved a whole extra redaction/rebuild cycle at the end.
Takeaway: Next time I'd paste the log and ask directly for exception logging around the ASR wrapper instead of first asking 'what's happening,' since that's exactly what we ended up doing anyway.
Takeaway: Next time I'd flag the .gitignore rule on .hackathon/submission up front so the agent checks tracked status before writing handoff files there, since I nearly lost those prompts silently.
Takeaway: Next time I'd ask for the eligibility audit and the improvement pass in one go, since the audit alone confirmed we already qualified and the real value was in the polish work.
Takeaway: I'd open with the HF Space compatibility question up front next time, since that's what actually gated the refactor — asking 'will this still deploy' before greenlighting the extraction was the right move.
Takeaway: I'd start the next prompt by pointing straight at the audio component's event bindings since that's usually where Gradio state gets stale, rather than describing the repro steps alone.
Takeaway: Next time I'd bundle the doc-consistency check into the original implementation prompt instead of asking for it separately, since the docs drift as soon as behavior changes.