Subscribe
Sign in
  1. Discussions
  2. Can you create a product without being a programmer, just...

Can you create a product without being a programmer, just by using AI?

Yuki Eliot
4 replies

Replies

AliThink
Gratitude
Gratitude
SmileAI is an AI-driven app designed to harness the power of AI to guide users in building positive emotions through daily questions and options. The app primarily leverages GPT and Claude on the front end, while the back end is powered by MarsCode, an AI plugin for VSCode. Even the product’s promotional static page was generated directly by Claude. https://smileai.alithink.com/ The overall development process relies on AI for nearly everything, with humans acting as the providers of the initial requirements. Here’s the general workflow: 1. Start with a basic product prototype for the app, then use multimodal large models to generate the core framework. 2. Iteratively improve the functionality and fix bugs based on the foundation. 3. Generate the necessary metadata for app deployment. 4. Take appropriate actions to address any review-related issues. @yuki_eliot
Share
Yuki Eliot
@alithink Oh wow, I need to learn from you. Are you on X?
Share
Launching soon!
Yes, it's definitely possible to build an MVP using only AI. This general process has worked for me: 1. Write specs and create mockups 2. Feed specs and mockups to AI 3. Implement AI code with back-end logic and ugly UI 4. Fix app logic bugs and API issues through testing and iterating with AI. (Some models' training sets include outdated API documentation. This can be addressed by directing AI to the latest docs.) 5. Provide additional prompts instructing AI to clean up UI. Iterate as needed. 6. Run automated security check on code I also suggest having AI write code for a framework with a lot of built-in utilities like Django, so you don't need to worry about creating admin, user authentication, and other housekeeping functions. A couple of caveats: 1. While you don't need to be a programmer to follow this process, you do need a general understanding of software development (e.g., what is an API, data persistence, etc.) and how the Internet works. 2. All bets are off when it's time to scale beyond the MVP stage. This is when human developers and systems architects who understand the risks/costs/tradeoffs of different computing approaches and back-end infrastructure choices will really add value. TLDR; Non-technical founders can and do use AI to code MVPs, but human programmers/systems architects are essential to scale. Hope this helps! 🙏
Share