Would you hire a VibeCoder to work on your product?

Hey Gabe, answering to your question: Yes and No

For frontend, I would hire someone who is smart in designing (UX/UI) but also knows how to revert the design into code.

For backend and things related to safety, I wouldn't hire a vibecoder assuming he/she has no knowledge in coding and how to ensure data security.

for sure. provided there's a good proof of work.

It entirely depends on how big the team is.

If it’s a team of 100+ people, I wouldn’t hire a VibeCoder.

But if it’s a company with just 1, 2, or 3 people= absolutely. If that vibe coder can build 50 times faster than other developers, why not? After all, that’s the fastest way to test my idea. If they can technically execute it in 5 days instead of 3 months, of course I’d trust them.

*Assuming they can deliver a quality product, not just slap something together.

Yeah sure but with an experienced with good fundamental on software development. Since when code based getting bigger it need foundation understanding especially when it come to debugging to fix the right problem.

Your question touches on a very relevant and nuanced topic in today’s engineering landscape.

As a senior engineer, I absolutely see the appeal of AI-assisted development and the "vibe coding" approach—especially in rapid prototyping, hackathons, or early-stage explorations. AI tooling, in particular, has unlocked a new level of speed and accessibility for builders of all levels.

That said, I believe there’s an important distinction to make between building fast and building well, especially in the context of long-lived, collaborative codebases.

Vibe coding—by its nature—often skips over crucial context: architectural decisions, edge cases, historical workarounds, and implicit contracts between systems. These aren’t always visible in code alone, and AI still struggles to reason across larger codebases with the depth and nuance needed to make safe, maintainable contributions. There’s growing empirical evidence that heavy reliance on AI coding agents in large or legacy systems can degrade quality, increase entropy, and introduce subtle regressions over time.

In my view, AI tools (and vibe coders) are best positioned as augmentation, not autonomy. They can assist in generating boilerplate, exploring variants of implementation, or even proposing refactors—but always under the guidance of engineers who deeply understand the domain and constraints. As you mentioned, the "core focus" should remain under the care of experienced developers who take long-term maintainability, security, and performance seriously.

Would I hire a "vibe coder"? Maybe—as a junior dev with strong product instincts and an eagerness to learn, paired with seniors who can mentor and gate their changes appropriately. But I wouldn’t rely on vibe coding as a primary development model in a production setting unless the problem space is exceptionally well-bounded, disposable, or non-critical.

In short: AI and vibe coders can be highly productive, but without guardrails, they can also be like pouring gasoline on a codebase fire. We should use them—but use them wisely.

@gabe Great thread, timely discussion. Wherever I do Fractional CMO consulting, I always insist on having 2-3 Marketers/ PMM with vibe coding skill. May not be required for Demandgen/ Inside Sales/ Inbound folks.

Don't have a product, but I'm of the same mind as @busmark_w_nika. Especially with security in mind, I would not hire a vibecoder unless they have some underlying knowledge, either in programming or security, hopefully both.

I'm not convinced by the hype around 'vibe coders'. At the end of the day, building a successful product still requires a solid foundation in software engineering fundamentals, not just a 'vibe'.

Meta just started allowing Vibecoding tools in their interviews! so it is "ZUCKERBERG" approved
and i think more and more companies will take this approach to ship faster!
So yaa why not!

Hell yeah I would. Bringing in a vibe coder make so much sense when you think of it form the product and user perspective. They bring a strong focus on what customers actually want. Their superpower or differentiator is knowing, communicating and translating a user's wants/needs into a viable product. "Knowing the vibe" if you will.

They're a new breed of PM.

It's a complimentary role that can elevate the product's overall appeal without pulling resources from dev.

Now would I trust them to do any of the backend, critical systems work? No. But that's why it's a team effort. I wouldn't trust a senior engineer who's focused on security to give me a product that users actually want to use. It takes two to tango!

Yeah, why not? Definitely 😄👍! As long as the Vibecoder has a solid grasp of programming fundamentals 💻. That way, they can maintain, understand, and fix any bugs that come up, especially from AI-generated code.

VibeCoders are the new wildcard, less about code purity, more about directional momentum. For teams iterating fast, they’re perfect for prototyping, vibe-checking UI tweaks, or pressure-testing ideas without slowing down the core loop. I wouldn’t hire one to own infra, but to explore offbeat angles? Absolutely.

good news: I am a competent programmer ( you'd be surprised at how competent I am given ( at the time of writing) I am 18.) I'd gladly take a job, on the sole proviso I am not required to work during times which conflict with school.

I’m planning to gradually scale this up.

Surprisingly, it’s been really useful to delegate repetitive tasks or minor fixes,much more than I expected.

So I’m going to continue expanding through experiments.

I don’t think this needs to be limited to just the early stage of a project either.

That said, one challenge is that code written by people without a technical background often takes a lot of time to clean up or refactor.

I’m still thinking about how to reduce that part.

I was amazed when a “vibe coder” delivered a working prototype in two days, freeing our core team for critical work. They’re perfect for rapid prototyping and minor features. Check out the Rutificador tool for personal and company data

why not

Hring a vibecoder makes a ton of sense if it’s for rapid prototyping, UI tweaks, or side features that need speed and experimentation. They’re great for shipping ideas fast without bogging down your main team.
When it comes to backend, data, or anything super important and could generate a risk factor, I’d still want someone with a strong fundamentals and real experience.

I am planning to consult early stage products as a "vibe coder". I am a senior engineer so I can guide the project but I was very skeptical when I started this journey a couple months back. I have posted about my experiments in multiple places but "proof of the pudding is in the eating".

So I am soon launching a vibe coded micro SaaS, sources available for anyone to learn from. https://github.com/brainless/letsorder and https://letsorder.app. Demo apps are going to be live soon but you can run it locally or check the commits, GitHub issues, etc.

I intentionally setup a Rust backend and TypeScript frontend (2 web apps) to take benefits of their type check systems. Also all API payload (JSON data) types are generated from Rust to TypeScript. These eliminate a lot of potential errors. I do not know any of the code, but I have set the initial instructions to use these technologies.