2 min lesson
Objections 3 & 5: the empirical reframe
Rebuild the parts of "Objections 3 & 5: the empirical reframe", then say why each one matters.
Step 1 of 2
Seniors hate AI / juniors stop learning both dissolve with one external data point: a University of Chicago study found more experienced developers were more likely to do an intentional planning step before generating code, especially with tools like Cursor. The skeptical senior's instinct - plan before you let it write - is exactly the best-practice the tool is built around. Cursor keeps the senior in the reviewer seat (Ask-mode explore, reviewable Plans, human-owned merge); it amplifies judgment, it doesn't replace it.
On juniors: AI collapses onboarding from a month-long process to days or weeks - new hires contribute fast partly because they can learn unfamiliar languages with AI. Ask modeA read-only mode for asking questions about a codebase without changing files; the safe way to explore unfamiliar or legacy code. Press Enter for the full definition. is itself a teaching tool ('explain this call graph'); Plans force juniors to articulate intent; reviews still happen. The risk isn't that juniors learn less - it's skipping the review gate, which you keep.
Why this matters mechanistically: plan-then-execute works because each model has limited context attention. Spending attention on planning, then breaking focused sub-tasks to task-specific models, lets each use its context window most advantageously. The senior's planning discipline is the same principle the harness runs on.
- What a request costs
- On legacy request-based plans a request ≈ 8 cents. Non-max models use ~1-2 requests; Max-mode models map their API cost back to a request count, which is how one Max chat can consume ~30 requests. Tokens map directly to dollars. This demystifies 'why did one chat cost 30 requests?' better than any list price.
- The dollars framing
- ROIReturn on Investment. The value gained versus what it cost, the language an economic buyer funds deals in. Press Enter for the full definition. is a journey: qualitative (are devs excited) → quantitative (PR velocityHow quickly pull requests are merged; the easiest delivery metric to measure, though it sits furthest from the customer outcome. Press Enter for the full definition., bug counts, % of code turned over) → dollars. The headline dollars framing is pull-forward revenue: as AI collapses deployment from weeks to days, teams pull roadmap items - and their revenue - forward. Cursor reports velocity up 30%+ and rising; automations save ~30-60 min/day per person.
Learn more
Full explanation
Full explanation
"You're right - and here's the part that changes the picture." Six words of concession buy you sixty seconds of genuine attention. Skip them and you're just another vendor arguing.
Don't over-promise on ZDRZero Data Retention. A contractual guarantee that the model provider won't store your code or train on it. Press Enter for the full definition.: zero-data-retention terms do NOT apply when the customer brings their own API keys. Get that wrong in a security conversation and you lose the whole room.
Older BugbotCursor's automated PR reviewer that posts inline findings and can push fix commits from isolated VMs. Press Enter for the full definition. pre-merge-resolution stats are directional; verify before quoting them live. Safer current facts: average review time is ~90 seconds, Bugbot finds ~10% more bugs per review and costs ~22% less per run.