Applied AI

Course map · S1 How LLMs work · Episode 10 of 12

Field note: reasoning versus standard

S1 E10 Field note · 20 slides · 7 October 2026

  1. Slide 1 of 20, Cover.
  2. Slide 2 of 20, Field note.
  3. Slide 3 of 20, Why it matters.
  4. Slide 4 of 20, Your turn.
  5. Slide 5 of 20, Finding.
  6. Slide 6 of 20, Method.
  7. Slide 7 of 20, Method.
  8. Slide 8 of 20, Method.
  9. Slide 9 of 20, Data.
  10. Slide 10 of 20, Data.
  11. Slide 11 of 20, Concepts.
  12. Slide 12 of 20, Data.
  13. Slide 13 of 20, Concepts.
  14. Slide 14 of 20, Data.
  15. Slide 15 of 20, Where it breaks.
  16. Slide 16 of 20, For engineers.
  17. Slide 17 of 20, For leaders.
  18. Slide 18 of 20, Decision rule.
  19. Slide 19 of 20, Open questions.
  20. Slide 20 of 20, Series map.

Lesson notes

Across three providers and 720 calls, high reasoning gained about 20 points on multi-step tasks and nothing on lookups or simple sums, mostly from one model.

Decision rule

Turn reasoning on for problems that need several steps held in mind at once. Leave it off for lookups and simple sums, and pilot before you commit.

Ideas in this lesson

  • Token cost
  • Reasoning
  • Measuring, not quoting
  • Latency

Take it further