{
"$type": "site.standard.document",
"bskyPostRef": {
"cid": "bafyreidsk2sws3vg7tqumzj4z2ec6tbocvieovh5fl46ntudborlib2soe",
"uri": "at://did:plc:pgryn3ephfd2xgft23qokfzt/app.bsky.feed.post/3mpkxoy5lic22"
},
"path": "/t/we-all-start-somewhere/177233#post_9",
"publishedAt": "2026-07-01T07:10:46.000Z",
"site": "https://discuss.huggingface.co",
"textContent": "Usually it’s the RAG setup, not the LLM. Refine your chunking, embeddings, retrieval, and prompt first. Retraining should be a last resort after you’ve confirmed the model is getting the right context.",
"title": "We all start somewhere"
}