Three frameworks — Caching (keep the question resident in the background), Radar (scan all stimulus through the question's frequency), Glasses and Hat (physically dress into research mode).
Anyone who's done research has had this experience. You pick a topic, ask AI 'research this for me,' and a minute later a clean summary appears with 10 links. And after reading, something feels off: 'this isn't research.' Information gathered; nothing connected. Why? Today we answer. The difference between search and research. Slowly.
Here's the spine. Research isn't information gathering. It's living with a question. Search ends in a minute. Research takes weeks, months.
Search and research look similar and get confused, but their essences are opposite. Search is 'finding the answer to a question you already know.' Research is 'carrying a question that has no answer yet.' AI is great at search. AI can't do research for you. Because AI doesn't 'live with the question' on your behalf.
Many people think AI made research easier. Information access sped up 10×. But in practice, something strange happens. Information piles up 5× more; the depth of the output actually shallows. Why?
Because the essence of research is not gathering but connection. 100 pieces of information that don't connect are just 100 pieces. There's no picture. Connections don't form automatically. They form when a question has been rolling in your head for a long time and, at some moment, piece A and piece B suddenly snap into each other. That 'suddenly' is the result of research. AI can supply information, but that 'suddenly' happens only in your head.
So 'time spent living with the question' is the essence. A minute isn't enough. Weeks at minimum, sometimes months. AI can't shorten that interval. AI can only keep supplying raw material through it.
Picture someone at the market picking a watermelon. A young person searches 'how to pick a watermelon': dry stem, dull sound when tapped, yellow on the bottom. Three facts in hand, they face 10 watermelons and can't tell which is which.
The grandmother is different. She's been buying watermelons for 30 years. The question isn't in her head anymore. The sense is in her body. She stands in front of the pile and picks one in 3 seconds. Ask why and she'll say 'just looked good.' She lived with the 'watermelon question' for 30 years. She sees them.
Research is the same. A good researcher has a topic 'stuck' to them. The topic surfaces in the shower, ads on the subway connect to it, a casual conversation drops a clue. The topic is in the body. In that state, AI giving 10 facts converts to depth instantly. Out of that state, AI giving 100 facts stays 100 facts.
Research runs on three frameworks: Caching, Radar, Glasses and Hat.
These three are today's core tools. Let's unpack.
1. Caching — the computing term for 'keep what you use near at hand.' In research: 'keep unresolved questions resident in the background of your brain.' Don't forget the question, keep it running. One thought before sleep, one after waking. 7 days, 14 days, 30 days. With that in cache, daily scenes start to collide with the question.
2. Radar — scanning surroundings at a specific frequency. With the question in cache, your brain filters for related information. Cognitive science calls this 'selective attention.' Same reason pregnant women suddenly see pregnant women everywhere on the street. Carry the question and its answers start jumping out.
3. Glasses and Hat — physically dressing into research mode. Surprisingly strong. Wear research-only glasses, go to a specific café, play a specific playlist, fix a specific time. Physical state signals the brain: 'this is question mode.' 30 days in, putting on those glasses drops your brain straight into research mode.
All three together and research runs in the background of daily life, automatically. AI then just supplies material.
Three stages.
1. Search — ask AI or Google, receive, done. 3 minutes. No depth.
2. Investigation — rummage through sources for a day or two, tidy up. 8 hours. Lots of information, no connections.
3. Research — carry the question for 2+ weeks while living. 100 hours. Connections and intuition emerge.
Many people mistake stage 2 for stage 3. Stage 2 never converts to stage 3 no matter how long you do it. Stage 2 is 'organizing mode,' not 'residing mode.' To reach stage 3, take the question all the way into the shower.
Let's do it. Question: 'How should education change in the AI era?'
Day 1 — write question on a note. Stick a post-it. First AI search — 10 points.
Day 3 — commute. Student in uniform. Flash: 'that kid is already using AI.' Jot in notes.
Day 7 — lunch with a friend, conversation drifts to kids' schooling. One sentence from them feels like a fragment of the answer. Jot.
Day 10 — shower. Sudden insight: 'teachers don't vanish — their role shifts.' Jot.
Day 14 — flip through notes. 14 days, 23 fragments. While reading, something strange happens. The 23 fragments connect into one structure. That's the 'suddenly.'
Then you ask AI 'write a research report with this structure,' and AI tidies it in 20 minutes. You don't use AI first. AI comes after your structure forms. Order matters.
Do these today.
Research differs from search. Search is 1-minute information retrieval. Research is the art of residing with a question for 2+ weeks. AI 10×'d the productivity of search. AI didn't shorten research. Connections happen only in your head.
If you want depth in the AI era, learn how to carry a question more than you learn how to use AI. Caching, Radar, Glasses and Hat. Plant these three frameworks into your life and AI gets truly useful — not as a thing that makes the answer, but as a thing that fills in the structure you built.
Caching → radar → glasses. Relearn research along these three.