Files

3.4 KiB

First Principles Examples - Artem's World

Example 1: TeddySnaps Photo System

Surface Problem

"We need to send more photos to parents"

Assumptions Challenged

  1. ❌ Parents want MORE photos → Actually: they want the RIGHT photos
  2. ❌ All moments are equal → Actually: moments with THEIR child matter
  3. ❌ Staff should curate → Actually: staff time is expensive and inconsistent

Fundamental Truths

  • Parents want proof their specific child is safe/happy
  • Face = identity = "that's MY child"
  • Time = context = "this happened today"
  • Automatic > Manual for consistency

Rebuilt Solution

AI face recognition + automatic tagging + parent-specific feed Staff takes photos freely → System sorts and delivers automatically


Example 2: TISA Enrollment Process

Surface Problem

"Our enrollment process is too slow"

Assumptions Challenged

  1. ❌ We need all info upfront → Actually: we need enough to make a decision
  2. ❌ Forms must be comprehensive → Actually: forms create friction
  3. ❌ Parents want to give info → Actually: parents want to RECEIVE confirmation

Fundamental Truths

  • Enrollment = exchange of commitment (parent commits child, school commits spot)
  • Minimum viable info: child name, age, parent contact
  • Everything else can come later
  • Speed of response = perception of quality

Rebuilt Solution

Phase 1: Name + age + email (30 seconds) → Instant spot reservation Phase 2: Detailed forms (after commitment) Result: 10x faster, higher conversion


Example 3: GolfTab Ordering

Surface Problem

"Golfers need to order food on the course"

Assumptions Challenged

  1. ❌ Need full restaurant menu → Actually: simplified menu works better
  2. ❌ Need precise delivery time → Actually: "by hole X" is enough
  3. ❌ Complex payment flow → Actually: charge to room/tab, settle later

Fundamental Truths

  • Golfer's hands are often dirty/gloved
  • Attention span on course = minimal
  • Location = hole number (already known via tee time)
  • Speed > options

Rebuilt Solution

5-tap ordering: Item → Quantity → Hole → Confirm → Done No menu browsing, no payment friction, no address entry


Example 4: Staff Scheduling (TeddyKids)

Surface Problem

"Staff scheduling is complex and time-consuming"

Assumptions Challenged

  1. ❌ Manager must create schedule → Actually: AI can draft, manager approves
  2. ❌ Staff preferences are unpredictable → Actually: patterns emerge over time
  3. ❌ Ratio compliance is manual → Actually: this is pure math

Fundamental Truths

  • Child:staff ratios are legally mandated (deterministic)
  • Staff availability has patterns (learnable)
  • Cost optimization is mathematical
  • Human oversight needed for exceptions only

Rebuilt Solution

Algorithm generates schedule → Manager reviews exceptions → Staff confirms 80% automated, 20% human judgment


Template for New Problems

PROBLEM: [What are you trying to solve?]

CURRENT APPROACH: [How is it done now?]

ASSUMPTIONS (list 3-5):
1. We assume [X] → But what if [challenge]?
2. We assume [Y] → But what if [challenge]?

FUNDAMENTAL TRUTHS:
• [Irreducible fact 1]
• [Irreducible fact 2]
• [Irreducible fact 3]

REBUILT SOLUTION:
[Solution built only from truths above]

VALIDATION:
- Does this solve the core need? [Y/N]
- Is it simpler than current approach? [Y/N]
- Can we build this? [Y/N]