Guillermo Rauch came by CS 210 the other week and demoed v0 to the students. It was a super cool talk and was just one sign of how AI has completely changed the way we prototype. The presentation was so awesome that I ended up rebuilding my personal website that very same weekend (which is where you’re now reading this post).

210 is a class that has traditionally focused on rapid prototyping, which includes starting with mock-ups and much more lightweight proof-of-concepts. However, the advent of generative AI has allowed for functional prototyping: rather than just simple frontends, we now can get fully-functioning applications with full-stack and ML capabilities through just a series of prompts. This doesn’t even cover other AI tools for development, which can help to supercharge the iteration process.

However, as we are able to get stuff done quicker, there’s another downstream consequence: the ability to more quickly receive feedback from users and tune our assumptions. Whether you’re working in the context of an established organization or are an early-stage startup on the way to PMF, having a prototype from scratch makes it even easier to test for desperation, solicit feedback, and code at a quicker rate. While true for consumer-facing companies — use opinion-based design and ship your ideas — I also think it’s beneficial for B2B companies that are finding design partners. Especially in the case where your startup might be building a net new capability that is only possible in the current state of technology (i.e., with LLMs/AI more generally), having a demo to showcase what is possible with your solution can only sell them further.

Long story short: build quickly, test quickly, and iterate quickly. The loop doesn’t change: it only accelerates.