1 min lesson
Models, modes and judgment
Use two rows in "Models, modes and judgment" to state the practical decision rules.
Step 1 of 2
Cursor explicitly prizes the judgment to reject bad AI output and an SA who can't teach that judgment is selling autocomplete with a bigger marketing budget.
Model selection comes up in every demo and most discoveries and the honest answer is that it's a trade-off, not a leaderboard. Faster models keep you in flow and cost less per call. More capable models reason through harder, multi-step work but cost more and run slower. Auto-selection routes the task to a reasonable default so most developers don't have to think about it.
- You optimize for
- Flow and quick edits
- Lean toward
- A fast model
- Real cost
- May stumble on genuinely hard, multi-file reasoning
- You optimize for
- Hard refactors and tricky bugs
- Lean toward
- A more capable model
- Real cost
- Slower and more expensive per task
- You optimize for
- Not having to choose
- Lean toward
- Auto-selection
- Real cost
- Less control on the rare task where the choice matters
| You optimize for | Lean toward | Real cost |
|---|---|---|
| Flow and quick edits | A fast model | May stumble on genuinely hard, multi-file reasoning |
| Hard refactors and tricky bugs | A more capable model | Slower and more expensive per task |
| Not having to choose | Auto-selection | Less control on the rare task where the choice matters |
Capabilities and names change; teach the trade-off, then confirm the current model lineup.
The exact models, prices and Auto behavior shift over time. Coach the trade-off framework and check the current lineup before quoting specifics.
Reasoning from the trade-off reads as more credible than reciting a list that may be stale by the interview.