#AI: Iteration > Prompt Engineering
The cult of the “perfect prompt” died young.
Modern LLMs —like ChatGPT or Mistral’s LeChat— have memory and keep getting better at understanding.
Just like in real life, the important thing is not to get it right the first time, but to iterate: add context, refine, and push the model to really think with you.
LiberLion
npub1wp…hzs7p
2025-09-08 09:58:44 UTC
Author Public Key
npub1wpzpvzfkn4m754fasp0wnt6ck20ycww4kz9nj4n5rquu9ul7a0xq4hzs7pSeen on
wss://nos.lolPublished at
2025-09-08 09:58:44 UTCEvent JSON
{
"id": "31ff98176f281783b43e075bfc6324a87a9188f72f1f51b575e9b961fb743bc0",
"pubkey": "70441609369d77ea553d805ee9af58b29e4c39d5b08b3956741839c2f3feebcc",
"created_at": 1757325524,
"kind": 1,
"tags": [
[
"alt",
"A short note: #AI: Iteration \u003e Prompt Engineering\n\nThe cult of t..."
],
[
"t",
"AI"
],
[
"t",
"ai"
]
],
"content": "#AI: Iteration \u003e Prompt Engineering\n\nThe cult of the “perfect prompt” died young.\n\nModern LLMs —like ChatGPT or Mistral’s LeChat— have memory and keep getting better at understanding.\n\nJust like in real life, the important thing is not to get it right the first time, but to iterate: add context, refine, and push the model to really think with you.\n\n",
"sig": "7b5b716b67dd25bf4154e56169489fc2521a90d9f4d6b8738946dc4c390299f7654da90865fc650a9c84ead8761427fba452dd123abc3d1f9730c5e57199aa45"
}