Bryan Hannes retweetledi

There are multiple way of using AI when it comes to code generation.
I built this app in minutes 👇
The first way is to let AI generate everything which results in:
☘️ You can generate anything but,
👉🏼 You burn a lot of tokens
👉🏼 You have no control of the output
👉🏼 The code is not according to your guidelines and best practices
The second way (shown in the video) is to use your own generators that use LLM's behind the scenes the fill in the blanks which result in:
👉🏼 You can't generate anything, it's limited to what your generators can do but,
☘️ You don't burn a lot of tokens
☘️ You have complete control of the output
☘️ The code is according to your guidelines and best practices.
Nitrokit uses NX generators that generate features that use LLM's behind the scenes for things the generator can't do by himself. I created this app for 22 cent in tokens, and it is completely according to my guidelines and architecture.
☘️ Authentication
☘️ Light/dark mode
☘️ Mobile friendly
☘️ Full crud functionality
☘️ Search
☘️ Backend is also generated with:
☘️ Many to many relationships
☘️ One to many relationships
☘️ Internationalisation (completely generated)
I'm not saying you should use Nitrokit and integrate it in your projects, but there are multiple ways of becoming crazy productive when it comes to #GenAI
Enjoy the video!👇
English







