Get in touch→
Photo · Kit (formerly ConvertKit) / Unsplash ← All insights
TalkOct 24, 202653 slides

How to validate your idea

Test fast, fail less, build smarter. A 90-minute session for the teams at the AI4Sustain makeathon, on finding out whether an idea is worth building before spending two weekends on it.

Building got cheap. A team can have a working demo in an afternoon now, so being able to build proves very little. What still decides whether an idea lives is whether someone needs it, and you only find that out by talking to real people and watching what they actually do.

The session moved through three parts: the idea (what problem are you solving), the customer (who you are building for), and the test (how fast and how cheaply you can find out). The teams answered along the way, with their own ideas on the table.

  1. Slide 1 of 53
  2. Slide 2 of 53
  3. Slide 3 of 53
  4. Slide 4 of 53
  5. Slide 5 of 53
  6. Slide 6 of 53
  7. Slide 7 of 53
  8. Slide 8 of 53
  9. Slide 9 of 53
  10. Slide 10 of 53
  11. Slide 11 of 53
  12. Slide 12 of 53
  13. Slide 13 of 53
  14. Slide 14 of 53
  15. Slide 15 of 53
  16. Slide 16 of 53
  17. Slide 17 of 53
  18. Slide 18 of 53
  19. Slide 19 of 53
  20. Slide 20 of 53
  21. Slide 21 of 53
  22. Slide 22 of 53
  23. Slide 23 of 53
  24. Slide 24 of 53
  25. Slide 25 of 53
  26. Slide 26 of 53
  27. Slide 27 of 53
  28. Slide 28 of 53
  29. Slide 29 of 53
  30. Slide 30 of 53
  31. Slide 31 of 53
  32. Slide 32 of 53
  33. Slide 33 of 53
  34. Slide 34 of 53
  35. Slide 35 of 53
  36. Slide 36 of 53
  37. Slide 37 of 53
  38. Slide 38 of 53
  39. Slide 39 of 53
  40. Slide 40 of 53
  41. Slide 41 of 53
  42. Slide 42 of 53
  43. Slide 43 of 53
  44. Slide 44 of 53
  45. Slide 45 of 53
  46. Slide 46 of 53
  47. Slide 47 of 53
  48. Slide 48 of 53
  49. Slide 49 of 53
  50. Slide 50 of 53
  51. Slide 51 of 53
  52. Slide 52 of 53
  53. Slide 53 of 53
1 / 53Tap a slide to look closely

01Six rules to remember

  1. A demo proves you can build, not that anyone needs it.
  2. Fall in love with the problem, not with your solution.
  3. Test the belief that could sink you, before anything else.
  4. Ask about the past, not the future.
  5. Behaviour beats opinion, and AI is an opinion.
  6. Decide what would prove you wrong, before you test.

02Keep the loop going

Believe, test, count, decide. Pick the belief that could sink you, run the cheapest test that could prove you wrong, count what people did rather than what they said, then persevere, change the scope, or pivot.

If you want one thing to do tonight: circle your riskiest belief and send three messages. Then move to the next belief, every week.

Presented at the AI4Sustain makeathon, 8th edition, Startup Garage @ TUIASI, Iași, Oct 24, 2026.

Written by

Ionut Maxim

Product design, strategy, and mentorship. Look carefully, say what seems true, move with intention.

Newsletter

Useful notes, occasionally.

Occasional writing on product judgment, design practice, and clearer ways to build. No noise.

Sent only when there is something useful to read.

Keep reading

Thinking · Jun 24, 2025Când toate semnele îți spun că n-ai timpAi observat?Thinking · May 20, 2025Cele trei întrebări care te salvează de la decizii proasteAcum câțiva ani eram într-o ședință din aia elastică. Stakeholderi importanți, timp puțin, și o decizie de luat: implementăm feature-ul X sau nu?Thinking · May 6, 2025Pierzi argumentul, faci level upEști supărat după o discuție în care celălalt ți-a dat cu tesla în perspectiva?

If this sparked a thought

The simplest next step is a conversation.

Get in touch→ All insights→
To IonutA thoughtWhat did this make you think of?Reading this, I thought about…I read every message myselfWrite →