I’m teaching CS177 Human-Centered Product Management in the fall. AI will be a core part of their work. But they need to understand what the advantage and the cost of it is. This is a lecture for my students who actually want to get hired.
I live in AI. I use Claude, ChatGPT, and Gemini daily. Because I write with AI, I know every quirk of every major LLM and can spot AI writing from across the room.
We will use AI in this class. But when I tell you not to use AI in this class, it’s not because I’m anti-tech. It’s because I want you to get hired and keep your job once you get it.
Employers Don’t Care About Your GPA
You’ve been taught success means getting a A, and it’s tempting to outsource your work when your uncertain. But employers don’t care about your grades. They care about whether you can think.
When you’re in that interview and the hiring manager asks, “How would you prioritize these features?” they want to see how your brain works. They want to watch you think through a problem in real time.
When your boss pulls you into a meeting about dropping retention numbers, you can’t say, “Hold on, let me ask ChatGPT.” You need frameworks and product intuition readily accessible in your head.
When you’re in the elevator with the CEO and they casually ask about last quarter’s product decision, your career trajectory gets decided by what’s in your brain.
How Learning Actually Works
Your brain has working memory and long-term memory. Working memory holds about seven pieces of information for maybe 20 seconds. Long-term memory is permanent storage.
Knowledge only becomes useful when it moves from working to long-term memory. That transfer only happens through effortful practice.
Every exercise in this class forces that transfer. When you struggle to apply a framework to a case study, that struggle moves knowledge from temporary to permanent. Skip it, and you’ll pass with a nice grade and zero useful knowledge.
MIT researchers proved this recently with brain imaging. They had students write essays over four months using ChatGPT, Google search, or just their brains. EEG scans showed ChatGPT users had 55% lower neural connectivity than students who wrote on their own. When tested later, 83% of the ChatGPT users couldn’t even recall key points from their own essays.
Students who struggled through writing themselves showed the highest brain activity and retained significantly more. When the ChatGPT group later tried writing without AI, their brains couldn’t activate the same neural networks—the “cognitive debt” had accumulated.
The struggle isn’t a bug in learning. The struggle is the learning.
Product Sense Is Compressed Experience
Product managers get promoted because they have good instincts. Junior PMs get fired because they don’t.
Product sense comes from seeing enough user research to recognize when something feels off. From working through enough prioritization exercises to spot hidden assumptions in roadmaps. From analyzing enough metrics to sense when numbers don’t pass the smell test.
You can’t get this from ChatGPT. These things only develop through repeated exposure to real problems and real analysis. Every shortcut robs you of experience that builds judgment.
The AI Learning Paradox
Using AI for coursework feels like learning. The output looks smart. You might even understand it when (if?) you read it.
But understanding someone else’s thinking isn’t the same as developing your own. When you ask AI to analyze a case study, you experience the endpoint of reasoning without going through the process. It’s like watching someone else do pushups and thinking you got stronger.
You pass with impressive-looking work and a brain that hasn’t been trained to do the job.
Why I Spot AI Writing
AI has tells. It loves certain phrases, structures arguments predictably, and hedges when it should be definitive. But more fundamentally, AI writing lacks the messiness of human thinking.
Real analysis has dead ends, reconsidered assumptions, breakthrough moments. It leaks into the writing. AI analysis is always clean and linear.
When you submit AI work, you reveal that you can’t tell the difference between sophisticated-sounding text and actual insight. We won’t mark you down because we can spot AI. We will mark you down because your reasoning is crap.
The Real Consequences
Use AI for this class and you might ace assignments and get decent grades.
Then you’ll bomb your first PM interview because you can’t think through problems without assistance. Or you’ll get hired and fired six months later when it becomes obvious you can’t do the work. Maybe, if you are lucky, you’ll muddle through as a mediocre PM for years, watching peers get promoted past you.
We’ll Use AI When It Makes Sense
I’m not asking you to avoid AI forever. We’ll use it in class when appropriate—to prototype interfaces quickly, vibe-check product concepts, or code when the learning objective is strategy, not programming.
But do your own thinking for the core learning objectives. Use these 10 weeks to build mental models you’ll rely on for your entire career.
When you get stuck on a tricky assignment, sit with that confusion. Work through it. Ask questions. Come to office hours. The struggle is the learning process.
You Chose to Be Here
This is an elective. You signed up for Human-Centered Product Management because you wanted to learn this job. You could have taken an easier class that required less work. You could be a social dance right now. You could be drawing naked people. But you are here.
Don’t waste that choice by taking shortcuts that rob you of the learning you came here to get.
The Bottom Line
You’re not paying for a grade. You’re paying to develop skills that make you valuable.
If you want the credential, use AI and get your A. If you want the capability, do the work and build your brain.
Your future hiring manager won’t ask for your transcript. They’ll ask you to solve a problem and watch how you think.
Make sure you have something to show them.