Building a Full Stack Application with AI

It is crazy to see how AI technology is shaping up. I have a theory that AI can be used to build a modern full stack application with just plain text inputs. Such as, "I need a CRM to manage my customers in the finance industry. I need to be able to edit the customers and move them to various stages."?

We should be able to generate a modern app with just that. And we are close to achieving it. Before getting into that, let's discuss what a modern stack is. Modern full stack applications comprise of three layers. API logic layer, Database layer, and Frontend layer.?

We could use AI to build a complete working system if it can diffuse the requirements and build cohesive components of those three pieces. Why would this work? GPT-3, which has changed how we consider AI services, has a great deal of knowledge about the internet, which includes programming languages. Codex, a model by OpenAI, can write and understand code. Based on my experience, one interesting thing is that even the normal GPT-3 models can spit out code. Especially good NodeJS code.?

Here is the interesting part. We managed to build a prototype that already helps us build backend DB schema and API layer with plain text input. Only the Front-end layer support needs to be added. Of course, you can't build the next Facebook, but a simple CRUD application, like a CRM and an employee tracking system,?with a full login/registration functionality, is a click away.?

The frontend layer can be the Angular/React framework. The developed codebase may need to be 100% working, but a single developer spending an hour can able to fix any glitches in the code.?

What does this mean? If we can build a full stack application from plain text, I see a few things happening,?

1. Reduced time to bootstrap a project. Since the basic data modeling, API layer, and UI components are built for you.?
2. Some users use low code solutions, like WordPress, only to come back again to a modern, customizable tech stack. They can now directly start with a modern tech stack.?
3. Developers only need to spend time on complex algorithms and integrations. No need to wade through authentication setup or DB modeling anymore.?
4. More people will build software tools as the entry barrier is low. Opening up for more innovative software tools.?

Not only this, GPT-3 can revolutionize many domains, including,?

• Customer support?
• Learning management systems (Such as personalized learning programs)?
• Research tools?
• Financial analysis?

I will probably write another post on that.?