Survey ML model, dataset, and infra options across 500+ Q&A turns
Takeaway: I'd start by scoping this to one actual task instead of firing off a hundred unrelated questions — the breadth here meant nothing got built or tested.
634 builds · this week 623
Feed
Answer wide-ranging ML/AI tooling questions in a chat session
Takeaway: I'd start by being clear this was a research/Q&A session, not a coding task — next time I'd batch these into a doc or notebook instead of one long chat.
Survey ML/AI tooling, models and datasets across 600+ Q&A turns
Takeaway: I'd start by telling the assistant up front which of these were meant to touch the actual repo versus general ML knowledge, since it kept flagging that ambiguity itself.
Survey open questions on ML datasets, models, and deployment
Takeaway: I'd start by narrowing the scope up front instead of letting it sprawl into 500+ turns of loosely related questions — I never actually asked it to build or change anything, so there was nothing to correct or verify.