I Built a Cover Letter Generator in 2 Hours. Is That Vibe Coding?

There's been so much talk about "vibe coding". I've watched over a dozen videos on YouTube about the term and I have read a couple dozen more articles on the topic. There seems to be more "anti-vibers" than "pro-vibers" out there. It's understandable, being that the idea was originally a long tweet from Andrej Kaparthy that's only a few months old. But the idea went viral and now there's an O'Reilly book about it.

I've tried taking a more pragmatic stance ever since we started getting the first tooling around code generation with LLM's and seeing people's reactions when CoPilot came out and then tools like Cursor. It's interesting because at first I hated the term "Prompt Engineering". Still kinda do, but I have accepted that that's what the tech zeitgeist decided to name the careful crafting of prompts to generative AI like LLM's to output something useful from it. It is a legitimate skillset to have and it's not as simple as asking questions to a chatbot.

Vibe coding on the other hand I'm not so sure yet. At least those precise words and the ethos of it. It's still too new, too vague of an idea, and I'm not so sure what makes it different from what we already have named as prompt engineering. I suppose it implies that there's less "engineering". There's less careful crafting of the prompts. The prompting is abstracted away from you and you're in less control of it. That said, I still haven't tried any of the newer tools like Claude Sonnet, or Windsurf, or the new OpenAi one. It seems like a new one comes out everyday.

But the reality is that anyone can now make apps much quicker. This is not something new, but now non-programmers and non-technical people can make more sophisticated software than say, just a Website with Wix or Squarespace.

This Sunday I made a Cover Letter Generator in maybe 2 hours tops. I don't know if it was technically "vibed", but I made all with ChatGPT 4o, talking to it like it was my experienced Web Developer buddy and copy and pasting pretty much all the code. This app seems like a trivial thing, something that's not serious, and not ready to be in "production". Well, it's hosted on Vercel and so far it's working as it's supposed to, with a Stripe integration and everything. (Btw if you want to show support and find the app valuable, or simply just like me, be sure to pitch in)

I don't know if the conclusion is that if you don't use these things you'll "get left behind", but I made something in two hours that people could potentially pay me $5 bucks for, and I'm just a dude without a programming job.

Sure, if all of a sudden I get hit by thousands of users I would probably need to think about "scaling" issues and add more dollars to the OpenAI token limit. But the point is that I've wouldn't have been able to make this as fast as I did and I would've probably got stuck on setting up Stripe or Googling issues with Tailwindcss.