WEEK 1

SE and What AI Changes

COSC 40943 · Senior Design · Week 1

WEEK 1

The question this course is about

What should software engineering become when AI is a permanent member of every development team?

Everything for the next fifteen weeks is an attempt to answer that on a real project, for a real client, with your name on it.

WEEK 1

Before I say anything else

Ask the room

Hands up: who used a coding agent this summer?

Who interned?

And the one I actually want: what is one thing you do not like about working with an agent?

WEEK 1

You already know how to program

That is the prerequisite, not the course.

Programming

Producing code that works.

Software engineering

Everything that makes it the right code, keeps it right as it changes, and lets six people who do not share a brain build it together.

WEEK 1

What is the "engineering" doing there?

The application of a systematic, disciplined, quantifiable approach to the development, operation, and maintenance of software.

IEEE Standard 610.12

Key point

The word was coined at a NATO conference in 1968, on purpose, because projects were failing and nobody could say why.

WEEK 1

Coding is one box

flowchart LR
  A[Vision] --> B[Requirements]
  B --> C[Architecture]
  C --> D[Design]
  D --> E[Code]
  E --> F[Test]
  F --> G[Release]
  G --> H[Maintain]
  H --> B
Key point

One box of eight. Agents are strongest at that one. You are accountable for the loop.

WEEK 1

The iron triangle

Scope features, functionality Quality Resources cost, budget Schedule time
Key point

Fix all three and quality is the only thing left to spend.

WEEK 1

What a requirement is

  • A capability the system must provide, or a constraint it must satisfy
  • Agreed with someone who can accept or reject the system
  • Verifiable: there is an observation that settles the question
Key point

"It needs to be fast" is a wish. "A dashboard page returns in under 400 ms at the ninetieth percentile with 200 concurrent users" is a requirement.

WEEK 1

You have read the headlines

  • Tech job cuts in the first half of 2026: 83% above the same period in 2025
  • Total US cuts across all sectors over that period: down about 40%
  • The entry-level rung is the part that thinned
Pitfall

You are about to graduate into this. Pretending otherwise would waste your semester.

WEEK 1

What actually happened

  1. The hangover. Companies hired through 2021 as if 2021 were normal.
  2. A convenient reason. Sam Altman calls it AI washing: blaming AI for layoffs you would have done anyway.
  3. A lost bet. Meta burned thirty billion dollars on Reality Labs in two years. Nobody laid off the metaverse.
Key point

AI is part of this. It is not most of it.

WEEK 1

The part that is about you

  • The tasks that used to train a new graduate (small specified changes, tests, boilerplate) are four-minute tasks
  • So the worry is not that AI replaces software engineers
  • It is that one senior engineer with agents replaces several juniors
Key point

Those junior jobs were how people became senior. Nobody owns the problem of replacing them.

WEEK 1

That said, the agents really are that good

SWE-bench: a real GitHub issue, a real repository, and a patch that has to pass the project's own tests. Verified is the 500-problem subset screened by hand in 2024.

  • 2023, on the original benchmark: about 2%
  • Today, on Verified: the large majority, solved
  • The top few models sit within about two points of each other
The agent

The benchmark is nearly saturated. That is the news, and we now have to go looking for tasks that are hard.

WEEK 1

Then they fall off a cliff

METR measures how long a task is, in human terms, and asks whether an agent can finish it.

Under four minutes of human work

Agents succeed on nearly all of it.

Over four hours of human work

Agents succeed on under ten percent.

Key point

Task length is the single strongest predictor of failure. Everything this course teaches lives on the right-hand side.

WEEK 1

So, honestly

Ask the room

Will software engineers be replaced by AI?

Say what you actually think. I am not grading this.

WEEK 1

Kent Beck did the math

Kent Beck created Extreme Programming and popularized test-driven development. He wrote this the first day he used a language model:

The value of 90% of my skills just dropped to $0. The leverage for the remaining 10% went up 1000x.

WEEK 1

The 10%, in his own words afterward:

  • Having a vision
  • Breaking that vision into milestones
  • Managing the design
  • Controlling complexity
Key point

This is the four-hour column. Beck said it two years before anyone measured it.

WEEK 1

The line I want you to leave with

Key point

You will not be replaced by AI. You will be replaced by someone who can use AI.

WEEK 1

AI is an amplifier

It multiplies whatever judgment you feed it.

  • A clear specification in: a great deal of good work, fast
  • Nothing in: nothing, faster
Key point

Zero multiplied is still zero.

WEEK 1

And the sign matters

Pitfall

A wrong premise in, and the wrong thing arrives complete, tested, documented, and spread through the codebase.

Worse than nothing, because undoing it now costs more than building it did.

WEEK 1

Execution, judgment, agency

  • Execution. Carry out a task someone already defined. Becoming cheap.
  • Judgment. Decide what matters, and catch what is wrong even when it looks right.
  • Agency. Own the problem. Take it from zero to one.
Key point

Execution is abundant. Judgment is scarce. Agency is the differentiator.

WEEK 1

Taste

Knowing what good looks like without needing a rubric.

  • Which abstraction will hurt in six months
  • Which test is theater
  • Which explanation is fluent and hollow
Key point

Taste is what lets you reject work. Without it you approve whatever arrives.

WEEK 1

Easy to look competent

Pitfall

AI makes it easy to look competent without being competent.

Polished output. Shallow thinking. Confidence growing faster than competence.

WEEK 1

What the job postings actually ask for

  • Code review, standards, source control, builds, testing, operations
  • Scoping requirements through launch
  • Communicating with users, other teams, and senior management
  • Mentoring, and influencing how a team works
  • Knowing when to adopt a technology and when to build one
Key point

Not one of them is typing.

WEEK 1

And now they ask for this

Postings in 2026 list AI-assisted development as a requirement, not a perk.

  • Named tools: Claude Code, Cursor, GitHub Copilot
  • Evidence you have shipped real work with them, not that you have tried them
  • Judgment about when not to use them
The agent

"Familiar with AI coding tools" is now the same kind of line "familiar with version control" was in 2010.

WEEK 1

The kitchen you are about to run

flowchart TD
  C["Executive chef: you"] --> M["The menu: spec and design"]
  M --> S1[Saucier]
  M --> S2[Poissonnier]
  M --> S3[Pâtissier]
  S1 --> P["The pass"]
  S2 --> P
  S3 --> P
  P --> V["Service"]
Key point

Nothing leaves the kitchen without crossing the pass. The pass is code review, and you are standing at it.

WEEK 1

You sign it

Key point

An agent can write it. You are the one who signs it.

  • The commit carries your name, the pull request carries your approval, the outage carries your explanation
  • The standard is reviewed and good, not "no AI"
  • Fluent output nobody read is slop, and the name on it is yours
Pitfall

The explanation test: if you cannot explain it, you did not review it, you forwarded it.

WEEK 1

Fifteen weeks to demo day

Six students. A real client. An agent on the team from day one.

Ask the room

Watch what happens to them, and hold one question: which of it would a better programmer have fixed?

WEEK 1

Nine things, four frames

The autopsy said You have a name for it now
Nobody wrote down what "done" meant. The client was there once. Features nobody asked for. A requirement is agreed and verifiable. Anything else is a wish.
No plan, so nothing could be behind. One person's absence stopped it. Quality is what got spent. Fix all three corners and quality is the only thing left to spend.
Six branches, one integration. Nobody read anyone else's code. "Almost done" was never checked. Nothing leaves the kitchen without crossing the pass, and you sign what leaves.
Every one of them arrived sooner and bigger. The amplifier does not check whether the premise was right.
Key point

Not one of the nine is a programming mistake, and every one of them has a name you learned this hour.

WEEK 1

This week

  • Today: the interest and skills survey. It closes Friday, and it is the only input to your team and your project.
  • Today: apply for the GitHub Student Developer Pack. Verification takes days.
  • Wednesday: the Napkin, and a live agent session that goes off the rails.
  • Friday: studio. Project Pulse running on your machine, plus one AI-assisted task on real code.
  • Next Wednesday, Sep 2: teams, clients, and project briefs.
Pitfall

Friday is not optional, the MVP is not optional, and "the agent wrote it" is not a defense.

WEEK 1

Three things changed

  1. Transformation got cheap. Anything whose difficulty was mostly typing is now nearly free.
  2. The cost of being wrong went up. Slow code used to catch bad requirements by friction. That friction is gone.
  3. The economics of rigor inverted. Practices were abandoned because upkeep cost human hours, not because they lacked value.
WEEK 1

Where the agent actually operates

flowchart LR
  A[Business need] -->|human| B[Requirements]
  B -->|agent drafts, human approves| C[Design of record]
  C -->|agent| D[Code]
  D -->|agent| E[Tests]
  E -->|human| F{Is this the right thing?}
  F -->|no| B
The agent

The agent is fluent on every arrow. It has no opinion whatsoever about the diamond.

WEEK 1

The delegation boundary

Stays human Goes to the agent
Requirements quality, scope, vocabulary Use case to design-of-record
Business constraints, domain knowledge Approved design to code and tests
Quality attributes, "good enough" Cross-document consistency checks
Decisions that are hard to reverse Mechanical transformations
"Is this even a good idea?" Drafting, refactoring, explaining code
WEEK 1

The test

Key point

If guessing it wrong would violate a requirement, the human pins it. Otherwise the agent derives it.

Both extremes are wrong answers. Delegate everything and you build the wrong product at speed. Keep everything human and you throw away the one advance that makes the rigor affordable.

WEEK 1

Your turn

Ask the room

Call out a task from your project. We place it on the boundary together, and you defend the placement.

WEEK 1

Why specs are the scarce thing

A programmer's spouse says: "Go to the store. Get a loaf of bread. If they have eggs, get a dozen."

He comes home with twelve loaves of bread.

WEEK 1

Ambiguity used to be survivable

Before

A vague requirement met a developer who asked a question, or built it slowly enough that someone noticed.

Now

A vague requirement meets something that never asks, never stalls, and produces two thousand confident lines before lunch.

Pitfall

The gap in a specification does not disappear when you hand it to an agent. It gets filled, silently, by something with no stake in the outcome.

WEEK 1

A session, in nine turns

You: Peer evaluations submitted after midnight save with the wrong week number. Find it.

Agent: Found it, the week is resolved from the server's local date. I can fix that. This class also has no tests and mixes date logic with persistence. Refactor it first?

You: Sure, that sounds cleaner.

Agent: Done, date logic extracted. While I was in there, the repository runs an N+1 query on every submission. Add a caching layer?

You: Go ahead.

Agent: Cache added, with eviction. This needs per-environment configuration. Set up the profiles?

You: Yes.

Agent: Profiles added. Note that Docker Compose does not pass the new variables yet...

Pitfall

Forty minutes. Six hundred lines. Three new abstractions. Evaluations after midnight still save the wrong week.

WEEK 1

Two ways a session goes wrong

Automation bias Agenda capture
What happens You accept output you did not verify You work competently on what does not matter yet
The tell You cannot explain a line you shipped You have stopped asking questions of your own
The move Verify against the spec, not the code Kill the session, do not redirect it
WEEK 1

The Napkin: six prompts

Twenty minutes, an unfamiliar problem, a defensible rough judgment.

  1. Shape: what kind of system is this really?
  2. The hard part: what makes it non-trivial?
  3. The bottleneck: what breaks first at scale?
  4. Stack: what would you build it on, and why that?
  5. Three kill risks: as mechanisms, not categories
  6. Verdict: feasible in the time you have?
Pitfall

Fixed order. Prompting the agent first destroys the exercise, because you cannot un-see its answer.

WEEK 1

Napkin round: the parking app

TCU Facilities wants a page showing students which campus lots have open spaces right now. Eleven lots, about 7,000 permit holders. Every lot already has entry and exit gates that count vehicles for the parking office, and those counts land as a CSV on a Facilities file share every fifteen minutes. There is no budget for sensors and the gates cannot be changed. They want it live before spring registration, and a Dean has already promised it in a newsletter.

Pitfall

Five minutes. Alone, silently. No agent, no neighbor, no phone. Six lines: shape, hard part, bottleneck, stack, three kill risks, verdict.

WEEK 1

What the agent said

Shape: a batch data pipeline with a thin read-only page on top, not a real-time system. Eleven rows of state, 7,000 read-mostly users.

The hard part: the gate counts are not occupancy, and the error compounds. Tailgating, motorcycles, propped gates, exits with no matching entry. And "right now" is a fiction: the freshest data is fifteen minutes stale, at exactly the hour it matters.

Bottleneck: trust, not throughput. One morning of "40 open" at a full lot ends adoption.

Stack: Spring Boot, Postgres, a small Vue page, Azure. Anything fancier solves a problem this system does not have.

Kill risks: data access never lands · a wrong number under a Dean's promise · scope creep.

Verdict: feasible, under-scoped if you count only the software. Cut "now" first: say "as of 8:45am, roughly 40 spaces (estimated)".

Key point

Thirty seconds, and it found the hard part.

WEEK 1

What scores, and what does not

Risk bingo

"Scope creep." "Technical debt." "Integration issues."

Nouns. Unfalsifiable. Zero points.

A mechanism

"The client's volunteer list is a shared spreadsheet, so two people editing it on Saturday silently overwrite each other."

Specific. Testable. Points.

WEEK 1

Where the detail lives

These slides are the fast version.

The SE and What AI Changes module has the full argument, the reading list, and the discussion questions.

SE and What AI Changes

space PgDn
Next step or slide
PgUp
Back
S
Skip to the next slide
O
Overview of all slides
N
Speaker notes
B
Black the screen
Ctrl+P
Print / PDF handout, one slide per page