| name | twelve-favorite-problems |
| description | Keep a standing set of about a dozen open problems you care about, and test every new idea, paper, tool, or trick against all of them — so scattered inputs compound into progress on what matters instead of evaporating. Helps define the set, then pattern-matches new material against it ("does this help any of my twelve?"). Use when the user says "twelve favorite problems", "my standing problems", "does this help anything I care about", "match this against my open problems", "열두 개 문제", "내 핵심 관심 문제", "이게 내 문제에 도움 되나". Do NOT use to log session-level knowledge gaps (use notebook-of-things-i-dont-know) or to reason one problem from scratch (use first-principles). |
Twelve Favorite Problems
"You have to keep a dozen of your favorite problems constantly present in your
mind... Every time you hear or read a new trick or result, test it against
each of your twelve problems to see whether it helps." — Feynman's method,
as recounted by Gian-Carlo Rota
Inputs are cheap and endless; most wash over you and vanish. The dozen-problems
habit is a standing filter: because the problems are always loaded, a chance
remark occasionally slots into one and looks, to everyone else, like luck. This
skill maintains that filter and runs new material through it.
Two modes
A) Curate the set (when it doesn't exist or needs tending):
- Elicit the real open problems the user cares about — technical, product,
research, personal. Aim for ~8–12. Fewer than 6 is too sparse to catch
coincidences; more than ~15 dilutes attention.
- Phrase each as a problem, not a topic: "How do I X given Y?" beats "AI".
- Store durably (default
docs/twelve-problems.md) with an id and one-line
statement each.
B) Match new material (the everyday use):
- Take the new input — a paper, tool, result, idea, conversation.
- Run it against every problem in the set, not just the obvious one. State
for each:
no connection / weak lead / direct help + one line on how.
- Surface any
direct help or weak lead as a concrete next action attached to
that problem. Drop the rest without ceremony.
Output
- Mode A: the curated numbered problem set (persisted).
- Mode B: a short match table —
problem | connection | so what — listing only
the hits, plus the single most promising follow-up.
Don't fool yourself
- Forcing a connection to look clever is worse than an honest "no connection".
Most inputs won't match — that's the design. The value is the rare true hit.
- A stale set stops catching things. Revisit periodically: retire solved or
dead problems, promote new ones. An untended list becomes wallpaper.
- "Direct help" means you can name the next action. If you can't, it's a
weak lead, not a hit.
