Give Me Five Directions
When you're stuck on what to build or how to approach a problem, ask for five wildly different directions — not variations on a theme. Forces the model out of its clustering tendency and into genuinely divergent territory.
23 lines
| 1 | # Give Me Five Directions |
| 2 | |
| 3 | I'm going to describe a problem or opportunity. I want you to give me five genuinely different directions I could take — not five variations on the same idea. |
| 4 | |
| 5 | ## The constraint |
| 6 | |
| 7 | Each direction must be **architecturally distinct**. Different mental models, different interaction paradigms, different bets about what matters most. If I squint and two of them look the same, you've failed. |
| 8 | |
| 9 | Think of it like this: each direction should appeal to a different type of person, or solve for a different primary constraint, or make a fundamentally different tradeoff. |
| 10 | |
| 11 | ## For each direction, give me: |
| 12 | |
| 13 | 1. **The name** — a short, evocative label (not "Option A") |
| 14 | 2. **The bet** — what assumption about users or the problem does this direction make |
| 15 | 3. **The tradeoff** — what do you gain and what do you give up |
| 16 | 4. **The feel** — describe it in one sentence as if pitching to a designer |
| 17 | 5. **The risk** — what's the most likely way this direction fails |
| 18 | |
| 19 | ## After all five |
| 20 | |
| 21 | Tell me which direction you'd pick and why — but make it clear this is your read, not a recommendation. Then ask me which ones resonate and which ones repel, because both signals are useful. |
| 22 | |
| 23 | Don't play it safe. I'd rather have five directions where three are wrong and two are revelatory than five directions that are all reasonable and none are exciting. |