Most developers trust AI to write a function.
But would you trust it with your entire project?
Multiple files.
Existing architecture.
Dependencies.
Business logic.
Cross-file changes.
Broken code that needs fixing.
That's a very different problem.
And that's exactly where I think AI development is heading.
NullCry is my attempt at building for that direction.
Instead of treating every prompt as an isolated coding task, NullCry is designed to work with the project as a whole.
Bring your own API key.
Choose your model.
Continue existing projects.
Make changes across files.
Fix problems across the codebase.
Generate complete projects.
I'm not claiming AI can replace developers.
I'm asking whether AI should be treated as a code generator at all.
Maybe the better model is an AI that works alongside the entire project.
Would you trust an AI with your codebase?
Yes or no — and why?
— Omar Zahouri
CEO of NullCry
Top comments (0)