Vibecoding
Build for the vibe, debug later
Start new thread
Alamedin Sabit

What AI project can I build solo in 30 hours? Need ideas!

by

Hey!

I want to dive into practical applications of generative AI and have set myself a challenge – to develop a useful product in 30 hours of focused work. My goal is not just an experiment but creating something with genuine practical value.

I have basic programming skills and can use any available APIs and tools (GPT-4, Claude, Stable Diffusion, etc.). The ideal project should:

- Solve a real problem

- Be feasible for one person to complete in 30 hours

- Leverage generative AI capabilities

- Not require training custom models

What would you recommend? Do you have personal experience creating such projects or ideas you've wanted to implement but haven't gotten around to?

I'm particularly interested in projects with monetization potential or those that could enhance a portfolio. Any advice on architecture, tools, or pitfalls you've encountered would be greatly appreciated.

Thanks in advance for your ideas!

157 views

Add a comment

Replies

Best
Hossein Yazdi

Have you asked AI? I think for scenarios as such, it's AI that's only helpful :)

Furqaan

Love the challenge.
One idea: Borrowed Brain - an anonymous voice-note platform where users submit 60-second problems, and get replies from 3 real humans offering their perspective (not advice). Full idea: https://ideatbd.com/p/so-that-did-not-go-as-planned

Why it fits:

  1. Real-world use case (mental clarity, decision support)

  2. AI augments flow but doesn’t replace human input

  3. Lightweight MVP possible in 30 hours

Build Strategy:

  1. Frontend: Simple form (HTML/React) with audio recorder + optional transcription using Whisper API

  2. Backend: Use Firebase or Supabase for auth, storage, and routing voice notes

  3. Distribution: Route submissions to a small pool of curated responders (initially friends or volunteers) via email or DM using tools like SendGrid

  4. AI Assist: GPT-4 for summarizing common patterns in replies, or flagging low-effort responses

Jihoon Lee

@chaosandcoffee Looks like a pretty cool idea, I guess the hardest part about building a platform like that would be maintaining the quality of responses on the platform :)

Furqaan

@indukeys Great point. One possible solution could be to start by onboarding a small group of vetted experts across key domains. Then, instead of an open Q&A free-for-all, we could implement a contribution-based matching system, where users earn credibility by giving valuable responses.

For example:

-Every helpful answer gets upvoted and adds to your “cred” score

-Higher scores unlock access to tougher questions or priority matching

-Askers can also “tip” responders (even just with shoutouts or badges)

-Over time, AI could help auto-route questions to high-signal responders in the right domain

Azfar

  1. Go to any keyword research tool

  2. Type 'AI generator'

  3. That's it, ideas will start flowing