So a ‘frenemy’ posted on LinkedIn about their vibe-coded NextJS ‘app’ that probably has just a simple SQLite database… but they managed to deploy it on Vercel - all after allegedly spending five evenings using Cursor.
So their app is super nothing interesting - they’re just excited about Cursor.
Thing is - I’ve been slogging away for weeks at something - configuring my developer environment, setting up Neon, my back-end, unit tests, E2E tests, component testing… just basically getting a skeleton together that actually deploys on fly.io and honing my Cursor rules so when I actually start building functionality, I’ll have a solid foundation.
Basically, this vibe-coding stuff is super bringing me downtown and I’m a bit like - screw this - shall I just vibe-code it too and sort out this stuff down the line?? I know the answer is no - but I’m getting cheesed off.
Hehe you can always try vibe-coding in a new project and just tell it to create an web app or so using general requirements. Its worth a try.
Some have managed to get quire far like the guy who made the airplane game but he is a dev too. Likely he researched first what libraries work well and easy even if not best solution overall. He had several iterations on it until it worked well once users had access.
Many (specially non-devs) get stuck and eventually it becomes a huge mess.
I wouldnt use that approach for anything serious.
Same as you mention setting up things makes sense and preparation is required.
For simple things we tend to overengineer solutions. Like if you need a one page solution then you dont need Postgres & Co, or NextJS
Sqlite is coming back big time. Can be used with lots of parallel users with correct settings. I would say that 90% of projects dont need massive super parallel handling capability and sqlite is a good solution. (not for all even on small scale, but still)
Vibe coding is nothing then clickbait these days.
Some people might get lucky results (dependent on the model, mostly nothing to do with Cursor, tbh).
The big difference: they got themselves an “app”, NOT a product!
After that first couple of iterations, you still have to put in actual brain cells’ steam and make it work (and hope you don’t have to remove half of the code from Sonnet-3.7).
It might be fun for a day, but I doubt many can get really far with it.
Yes not just that but for anything serious the whole thing would need to be properly rewritten.. So as proof of concept yes, for marketing perhaps too. but there it ends
Simple answer: Use your creativity, learn things and keep stacking from there complete some courses to get along with the commands and other components, modules, etc.
SQLite never left, it’s just that for the last 10 years you’ve been bombarded by youtube ads and conference bootlickers to promote whatever AWS wrapper was making them more money. Meanwhile SQLite is free, fast, and scalable, it lives in cars and on each Android App. Unless your data is highly relational and each query requires bajillion megabytes of indexes, the OS disk cache will make your DB access instant.
Vibe-photoshooting and vibe-videorecording were making money for people I knew because they had better business sense or paid connections, while using cheaper hardware and lower skills than I was.
It’s not about the skills, or tools, it’s about the vibe. It’s in the name. Get your vibe up! Deploy ZERO CODE or CODELESS (manually written) apps using vibe development today!
And then we have an @Cursor where we can throw ‘what_ifs’ at
GIVE ME ALL THE X ABOUT Z in a sqlite db and a localhost with graphs and tables and focus on Oligachs and table this that and the other thing" GO BI z N ITCH
Well there are two sides to it imo: the wanting to be technical approach for obvious reasons, then the vibe side, where at the end of the day (and using an analogy) do you really need a drill, drill bit and wall plug, or do you just want the picture on the wall?
