I created H.L.D.R to develop a product (or proof-of-concept) that harnesses GenAI for positive impact.

I often found myself becoming overly reliant on tools that were diminishing my confidence and inducing a sense of imposter syndrome in my work and knowledge.

It became increasingly difficult to tackle challenging problems without resorting to ChatGPT for solutions.

H.L.D.R was my solution to break this dependency.

The implementation of H.L.D.R follows a streamlined three-step process:

Step 1: The user writes their code as usual, focusing on solving a problem, creating a feature, or exploring algorithms.

Step 2: The file is saved. This can happen manually (via CTRL+S) or automatically based on the user's preferences. Saving acts as the trigger for the analysis.

Step 3: H.L.D.R produces insights directly in the command line. This feedback may include error detection, suggestions for optimization, or alternative solutions tailored to the user's coding style or goals.

While this approach is functional and efficient, it leaves room for improvement to enhance user experience and accessibility.

This current implementation has notable drawbacks:

• Not User-Friendly: A command-line interface may be intimidating for beginners and lacks visual clarity for quick feedback interpretation.
• Restrictive Platform: Currently reliant on GitHub Codespaces, limiting accessibility for developers using other environments or local setups.
• Limited Integration: Does not seamlessly integrate into popular IDEs or font editors like Neovim, VSCode, or JetBrains tools, limiting its usability across diverse workflows.
• Accessibility Hurdles: Requires manual setup and CLI familiarity, posing barriers for less experienced users.
• Absence of Real-Time Feedback: Insights are only generated upon saving the file, missing the benefits of live linting or error detection.

To address these challenges, I have open-sourced the project and proposed the following solutions:

• VSCode Extension: Develop a dedicated extension that integrates H.L.D.R into VSCode, offering a side panel for insights, inline feedback, and customizable triggers for analysis.
• Neovim Plugin: Create a plugin for Neovim leveraging Lua, featuring asynchronous communication, floating windows for insights, and configurable key mappings.
• Expanded IDE Support: Build plugins for other popular IDEs like JetBrains IntelliJ, PyCharm, and Sublime Text to make H.L.D.R widely accessible.
• Platform Independence: Provide a local installation option or containerized setup (via Docker) to reduce dependency on GitHub Codespaces.
• Improved Visualization: Enhance feedback display with Markdown-style formatting in CLI or graphical interfaces in IDE plugins.

Now provided the technical drawbacks, I still believe in the proof-of-concept and believe that this stands as a basis for the good that GenAI can do for to assist the current & future state of tech-talent.

Enjoy the video!