MS&E 275: AI-pilled
building the AI-native company
01what is this class about?+
For over a decade, we have taught classes in which we asked student entrepreneurs to build and test their hypotheses with real customers. While that core remains for every startup, almost everything around it has changed. Building a product used to be costly and risky. Now it costs close to nothing, so the product build is no longer a key risk. It’s merely the starting line. The hard questions have moved:
- Can any advantage last when a competitor can ship the same thing next week?
- How do you find and reach customers whose own tools, workflows and buying habits are being rewritten by AI?
- When you can build nearly anything, what do you choose NOT to build? What experiments do you choose NOT to run?
In practice, the founders who pull ahead know where they are going and why. This acts like a magnetic field that aligns all of the individual vectors whether they are agentic or human so that their efforts add up. Building is now cheap, so the real work is curating and refining: deciding what to cut, and making what's left usable and scalable. And because building is cheap, the why matters more, not less: it's easy to build something coherent and soulless in a week. We ask the why early and help you refine it.
This is not a class to train you to be a VC. It is not a class to convince you to found a company nor is it an incubation. Some students will build a venture-backed company; some a $10M business, a nonprofit, or a feature. All of them leave knowing how to use AI to build what is in their head and then how to evaluate what they have built.
The question every week: what did you build that you're throwing away?
- when
- Tuesdays, 3:00 to 5:50 PM · Jan 5 to Mar 9, 2027
add to my calendar: google calendar · apple / outlook (.ics) - where
- Y2E2 111
- format
- lecture, a guest, team share-outs
- instructors
- Ann Miura-Ko, Ryan Caldbeck
- speakers
- Guillermo Rauch (Vercel), Akshay Kothari (Notion), Seb Goddijn (Yak Labs, built Glass at Ramp), Matt Swulinski (Viktor), and more
02what are we looking for?+
People with an idea in their head and the drive to make it real.
- An idea, and a why. Something you want to exist, and a reason you're the one to build it. Venture scale not required.
- A record of executing. Something exists in the world because of you, and you carried it past the hard part.
- AI fluency. You already build and run things with AI, not just search with it.
- Judgment over output. You're as interested in what to cut as in what to make, and willing to change your mind on evidence.
About 30 students, solo or in pairs. Plan on 15+ hours a week. Each week you post a journal entry: what you built, what you changed your mind about, and what you threw away.
- applications close
- Mon Nov 30, 2026
- decisions
- by Thu Dec 3
03apply+
This class is a ten-week build. You arrive with an idea you want to make real, and you spend the quarter building it into a working AI-native company, alone or with one partner. The class meets once a week as a three-hour studio with the real work happening between sessions.
We have a short application to the class for two reasons:
- We want to prioritize students who have an idea going into the class and want to use these three units as the opportunity to build it out.
- The class prerequisites include having a clear project/company vision, a track record of successful execution, and demonstrated proficiency with AI tools. We're using this form as the opportunity to get a sense of those three things in order to make sure all of our students run equally as fast towards the goal line.
Every question is required, and the application should take around 30 minutes. We're not looking for polish; show us your idea, one thing you've shipped, and how you actually use AI.
Short, honest answers are preferable. While most of the class is about learning to use AI efficiently, we hope that your answers to these application questions can be written by you.
Your progress saves in this browser as you type.
Got it. Your application is in.
Decisions go out by Thu Dec 3. Until then: what did you build this week, and what did you throw away?
First class: Tuesday, Jan 5, 2027, 3:00 to 5:50 PM PT in Y2E2 111. Add this to my calendar: google calendar · apple / outlook (.ics)