experiment

Prompt loops as a study habit

Treating a model like a sparring partner — tight loops, written conclusions, no endless chat.

Hypothesis: short, deliberate prompt loops beat long freeform chats for learning retention.

Method

  1. State one question in a single sentence.
  2. Ask for a counter-argument, not a summary.
  3. Rewrite the answer in my own words offline.
  4. Log what changed in my mental model.

Early read

The forcing function is step 3. If I cannot rewrite it, I did not learn it — I only consumed it.

Q → challenge → rewrite → log

Next: run this for a week on one topic (embeddings) and compare notes density to unstructured chat.

← Field notes