I attended my first vibe coding hackathon (honestly, my first hackathon ever) on Saturday. I know I'm beating a dead horse when I say everything is changing by the hour, but it is hard to keep up. That said, it's also never been easier to try something when you're curious. Got a crazy idea? Just build it. Test it. See what happens.
I was dreading vibe coding because I thought I'd drown in technical jargon. But here I am, posting about it. Four takeaways, in increasing order of how much they've rearranged my thinking.
1. Prototyping is terrifyingly easy now
I think I can go from PRD to working MVP in 4–5 hours for a single feature. Not polished, but good enough to communicate the idea and test fast. I actually built two apps at the same time. (We even started a third, but pivoted once we realized it needed more time.)
Sit with that number for a second, because it changes the economics of product decisions. The traditional cost of validating an idea was high enough that we built entire disciplines around avoiding the build: opportunity assessments, prioritization frameworks, stakeholder pre-alignment. All of that existed partly because engineering time was the scarcest resource in the building. When a testable prototype costs an afternoon, some of those meetings should become prototypes instead. Not all of them. But more than we're currently comfortable admitting.
The catch (there's always a catch): a prototype that took four hours looks deceptively like a product that needs four months. The skill isn't building the prototype anymore. It's being honest about the distance between the demo and the production system, especially in environments where the production system has to handle compliance, edge cases, and angry customers. I've shipped AI in a regulated industry; the demo was never the hard part.
2. I've seen it and now I believe it
The challenge rules required the code to be 90% prompt-driven. We vibe-coded everything, even the backend — connecting Supabase to LLMs to power the app. Not production-grade, but excellent for validating ideas.
I'd read the breathless posts about non-engineers shipping apps and quietly filed them under "probably demos." Watching it happen in the same room, done by people like me, moved it from rumor to fact. The gap between "I have an idea" and "I have a thing you can click" has effectively collapsed. What hasn't collapsed is the gap between "a thing you can click" and "a thing a business can depend on." Knowing which side of that line you're standing on is becoming a core product skill.
3. The PM ↔ engineering gap is vanishing
Yes, this horse is also dead, but here we are. I asked a dev how he felt about building this fast. He shrugged: "We already use Cursor at work." Another said: "We only hire product engineers now."
"We only hire product engineers now" deserves more attention than the shrug it was delivered with. If engineers are moving toward product, the pressure on product managers is to move toward something — and I'd argue that something is judgment: which problems deserve solving, what "working" actually means, what breaks at scale, what regulators will say. The PMs who treated engineering as a black box could coast when the black box was expensive to open. Now that anyone can open it, "I don't do technical" is an expiring excuse.
I'm excited… and a little scared. It feels like a tipping point, and I don't know which way the future rolls. But I'd rather be the PM who has built things with these tools than the one reading trend reports about them.
4. Taste is the differentiator
When it's this easy to go from idea to product, what makes you stand out? The effort you put into making it seamless and delightful. The moments where you don't settle for the AI's generic fifth-draft output. AI won't break out of boilerplate on its own — not yet. That's still on you.
Everything the AI produces by default is, by definition, what everyone else's AI produces by default. The floor has risen dramatically; the ceiling hasn't moved. What fills the space between floor and ceiling is taste: knowing what to cut, what feels off, which of the five generated options is actually good and why. Taste used to be a nice-to-have on top of execution skill. When execution is nearly free, taste is the skill.
So what am I doing about it?
Practically: I keep building. The side project I started at this very hackathon grew into Momentum, a working app with three AI modes. Not because the world needs another productivity tool, but because opinions about AI products formed while building one are worth ten formed while reading about them. If you're in product and haven't spent a weekend making something with these tools yet, that's the single highest-leverage thing you can do this month. Come tell me what surprised you.