0 reactions
Frustrations with AI Assistants for Complex Coding Projects
As a developer, I often find chatGPT and other AI coding assistants fall short when working with larger, multi-file projects. The need to manually upload each file is cumbersome. File size limits constrain real-world codebases. And the lack of persistent memory across sessions means having to constantly re-explain context.
What if there was a tool that could ingest your whole code repository in one go? Not just as plain text, but in a way that understands code structure like functions, variables, and relationships. It would maintain memory on all those details between sessions, so prior context on a project is never lost. And it would integrate right inside existing IDEs like VS Code for a seamless developer experience.
This context-aware, stateful foundation could enable more intelligent suggestions, error checking, and autocomplete as you code. So rather than just a chat window, it acts like a supercharged AI pair-programmer that learns your code.
Before embarking on building something like this, I'm curious:
* Are there pain points with chatGPT and other assistants I'm missing? What has been your experience?
* What other capabilities would make a tool like this more useful?
* How could it fit best into your development workflow?
#technology
As a developer, I often find chatGPT and other AI coding assistants fall short when working with larger, multi-file projects. The need to manually upload each file is cumbersome. File size limits constrain real-world codebases. And the lack of persistent memory across sessions means having to constantly re-explain context.
What if there was a tool that could ingest your whole code repository in one go? Not just as plain text, but in a way that understands code structure like functions, variables, and relationships. It would maintain memory on all those details between sessions, so prior context on a project is never lost. And it would integrate right inside existing IDEs like VS Code for a seamless developer experience.
This context-aware, stateful foundation could enable more intelligent suggestions, error checking, and autocomplete as you code. So rather than just a chat window, it acts like a supercharged AI pair-programmer that learns your code.
Before embarking on building something like this, I'm curious:
* Are there pain points with chatGPT and other assistants I'm missing? What has been your experience?
* What other capabilities would make a tool like this more useful?
* How could it fit best into your development workflow?
#technology
| 18,000 mlx | |
| 18 | |
| 0 |