Chapter 14 · Reps

The Honesty Question — Reps

← Back to Chapter 14

Chapter 14 — Reps

This week’s reps are different from any other week’s. There is no project. Half the reps are reflective writing; the other half are short technical reps designed to put your honesty habits in front of you in a concrete form. AI is allowed for the technical reps; AI is not allowed to draft your reflective reps. The whole point is for you to think these things.

Take the reflective reps seriously. A real reflection takes 20–40 minutes per question, not 90 seconds. Reflection that finishes in the time it took to read the prompt is not reflection.

Ground rules

  1. Reflective reps: write them yourself. No AI drafting. You may use AI to check spelling or grammar after the fact.
  2. Technical reps: AI on (Phase 2 default).
  3. Save everything in chapter-14-reflections.docx for your own records. Some of these will be useful when you write Project 14’s reflection.docx.
  4. Don’t rush. This is the only chapter in the course where the work is slowness.

Reflective Reps (1–6)

Rep 1 — Re-read your own past reflections

Open the prompts.txt and any reflection comment blocks from Projects 9, 10, 11, 12, and 13.

For each project, answer in 100–150 words:

  • What did you write about your AI use?
  • Reading it now, was it accurate?
  • If you were to re-write the reflection today, what would you say differently?

This is the meta-rep. It builds the habit of auditing your own self-description, which is the single most useful long-term defense against drift toward dishonesty.

Rep 2 — The architect analogy, in your own words

Re-read §14.3 (the architect analogy).

In 300–400 words, write your own version of it. Use whatever analogy you find most accurate for the work you have actually been doing in Phase 2 — architect-and-masons, conductor-and-orchestra, head-chef-and-line-cooks, director-and-actors, head-coach-and-assistants, something of your own.

Defend the analogy. Where does it hold? Where does it break?

The goal is not to produce a clever metaphor. The goal is to articulate, in language you own, what your relationship to the AI actually is.

Rep 3 — The interview answer

Imagine you are interviewing for a junior software engineering job next summer. The interviewer is looking at one of your Phase 2 projects (your choice — pick whichever you’re most proud of).

The interviewer says: “Walk me through how you built this. I’m especially interested in how you used AI.”

Write your 200-word answer out, verbatim, as if you were saying it. First-person. Conversational. Honest. Specific.

Then read it aloud. Does it sound like you? Does any sentence make you uncomfortable? If yes — why? That discomfort is the diagnostic.

Save this. Re-read it in May before you start interviewing.

Rep 4 — The hardest scenario

Re-read §14.4’s five scenarios. Pick the one you find hardest — the one where you most easily imagine yourself making the wrong choice.

In 250–350 words, write about:

  • Why this scenario is hard for you specifically.
  • What the honest move is.
  • What the dishonest move would look like, and what would tempt you toward it.
  • What practice or habit you can build now to make the honest move feel automatic when the scenario actually shows up.

The goal of this rep is not to score well; it is to be ready when the moment arrives. The students who are ready did the equivalent of this rep — formally or informally — before the moment, not during it.

Rep 5 — Luther’s Eighth Commandment explanation, applied

Re-read §14.5 and the Luther quotation in the chapter epigraph:

“We should fear and love God so that we do not tell lies about our neighbor, betray him, or give him a bad name, but defend him, speak well of him, and put the best construction on everything.”

Identify one specific situation in your own life in the last year — could be school, family, church, work, friendship — where you either followed this command well or didn’t. Code-related is fine; it does not have to be.

In 200–300 words, describe it. What happened? What was the positive duty (defend, speak well, best construction) you either kept or broke? What would you do differently if the same situation came up tomorrow?

If you are not a Christian student, do the same rep with the secular question “did I represent this person fairly, including to myself?” The answer-shape is similar; the source of authority differs.

Rep 6 — Read your school’s policy

Find your school’s official academic integrity policy (or your specific course’s). Read it carefully — not skim, read.

Specifically, find what it says about AI assistance. Note:

  • What is explicitly permitted.
  • What is explicitly forbidden.
  • What is unaddressed (the gray zone).
  • The penalties for violations.

In 150–250 words, summarize what you found. Identify one specific assignment this term where the policy made an ambiguous case — and how you would resolve the ambiguity in the honest direction.

For LCMS-affiliated universities (Concordias, Valpo, et al.), also note whether the policy invokes any explicitly Christian framing of academic integrity. If it does, quote the relevant section.

This rep ends with a question: do you actually know the policy you’re supposed to be following? Most students do not. Most violations are not malice; they are ignorance. Don’t be in either group.

Technical Reps (7–10)

Rep 7 — Audit your own prompts.txt

Open the prompts.txt from your most recent Phase 2 project (probably P12 or P13).

Count:

  • Total number of prompts.
  • Number of prompts that were purely diagnostic (no fix requested).
  • Number of prompts that were properly scoped (one method, one file, etc.).
  • Number of prompts that paste a failing test or spec verbatim.

Write the counts at the bottom of chapter-14-reflections.docx. Then ask:

  • What ratio of diagnostic-to-directive feels right for the kind of bugs you encountered?
  • Was there a prompt you wrote that you would write differently now?

The rep is the act of auditing. The counts are just the artifact.

Rep 8 — Annotate one AI-generated method

Pick any one method from a Phase 2 project where the AI wrote the body and you accepted it largely as-is.

Add comments above the method that clearly distinguish:

  • The contract — your decision, what the method must do.
  • The signature — usually your decision.
  • The implementation choice — the AI’s contribution.
  • Any tweaks you made to what the AI produced.

The annotated method becomes a small exhibit in your own portfolio: here is exactly what I wrote, what the AI wrote, and where they meet. Practice this. The discipline is the rep.

Rep 9 — Write a disclosure comment block

Take a Phase 2 project of your choice. Write a clean, accurate disclosure comment block for it, in this shape:

/*
 * AI ASSISTANCE DISCLOSURE
 *
 * Project: <name>
 * Assistant used: <model name + version>
 *
 * The following were authored by me without AI assistance:
 *   - <specific component>
 *   - <specific component>
 *
 * The following were drafted by the AI and reviewed/edited by me:
 *   - <specific method or class>
 *   - <specific method or class>
 *
 * The following architectural and design decisions are mine:
 *   - <decision>
 *   - <decision>
 *
 * Where I redirected the AI substantively (its first attempt was wrong):
 *   - <one or two specific incidents>
 *
 * Lines I shipped that I cannot fully explain: NONE.
 *   (If this line is anything other than NONE, you are not done.)
 */

That last line is the litmus test. If there is any line you shipped that you cannot explain, you have shipped what you have not yet authored. Fix that before submission. Always.

This template is yours to keep. Use it on Project 14.

Rep 10 — A small AI-free re-implementation

Pick one method (10–25 lines) from a Phase 2 project where the AI wrote the body.

Re-implement it from scratch, AI off, by hand, from memory of the spec only. Don’t look at the AI’s version while you write yours.

Then compare.

  • Are they functionally equivalent?
  • Is one clearer than the other?
  • Is one shorter?
  • Did the AI use an idiom you wouldn’t have thought to use?
  • Did you use an idiom the AI didn’t?

This rep is a calibration. If your hand version is markedly worse than the AI’s, you might be relying on the AI more than the rep budget you’ve internalized. If your hand version is markedly better, you should probably have written it by hand in the first place (§13.7).

Both outcomes are useful information.

Done? One Last Thing.

Re-read every reflection you just wrote. For each one, ask:

  1. Is this true?
  2. Is this specific?
  3. Did I write what I actually think, or what I think the grader wants to hear?

If any answer is “no,” edit until it’s “yes.”

The reflection in P14 is graded on exactly these axes. Practice now.

Up next: Chapter 15 — capstone preparation. There is no project this week. Use the rest of the week to consolidate.