When to Wait vs When to Build Framework
A simple decision framework for rapidly, AI based, changing society.
1. Is tech improvement outpacing your execution?
If AI capability is growing faster than you can build manually, waiting may be optimal.
Indicators
- monthly frontier jumps
- new model releases invalidate work
- tools automating what you’re building
- your roadmap looks manual + labour heavy
→ Lean toward waiting
2. Are you capturing something that compounds over time?
Some assets compound regardless of tech shifts:
- audience
- brand
- trust
- governance position
- community adoption
- partnerships
- data you have permission to use
- distribution channels
→ Build these immediately
3. Is there a land-grab or a standards moment happening?
Early movers win when:
- standards are forming
- governance rules are being written
- ecosystems are emerging
- protocols and norms are still flexible
→ Move now (even if technology isn’t ready)
4. Will your work be leapfrogged soon?
Classic “Wait Equation” logic:
- if waiting 6 months makes the job 10× easier,
- do not build the hard version today.
Ask: Will this be trivial later? If yes → wait.
5. Can you build a thin layer now?
Instead of a big build:
- MVP
- integration layer
- proof-of-concept
- tiny demo
- community test
Build something extremely small → keeps you in motion without sinking cost
6. Reverse question:
What becomes harder if you wait?
Things that get harder later:
- trust
- coordination
- community memory
- legitimacy
- governance seats
- reputation
- distribution relationships
→ Build these now
Cheat-Sheet
| Item | Strategy |
|---|---|
| Tech Implementation | Likely wait |
| Governance / Standards | Build now |
| Community | Build now |
| Partnerships | Build now |
| Data Rights | Build now |
| Manual Tools | Wait |
| AI-Automatable Work | Wait |
| Distribution | Build |
| Brand | Build |
| Proof-of-Concept | Small now |
Quick-Rule
If AI will do this better than you soon → wait.
If humans and relationships are the advantage → build.
The Travelling to Mars Analogy
If you want to go to Mars, and every year rockets get ~10% faster, should you:
- launch now (taking longer), OR
- wait a few years and travel much faster?
If you leave now, a future mission might actually pass you on the way and arrive first.
This creates a weird implication: sometimes waiting gets you there sooner.
T = W + ( D / ( S(t + W) ) )
- W = how long you wait for better tech
- D = task size (distance to Mars, capability gap, etc.)
- S(t + W) = speed/capability you’ll have at time t
- T = total arrival time