Tools I Use
Informational
Overview
This is a curated list of tools I use for research, projects, learning, and other tasks. I use the latest models from these platforms as they become available.
AI 🤖
I regularly switch between different AI platforms depending on the task. Each has its own strengths and capabilities. Often I will run hard problems by multiple models or have one critique the output from another.
Coding:
- Claude Code - My primary coding tool
- OpenAI Codex - An incredibly capable model/harness and great for long-running and complex tasks. Also a great code reviewer.
- Gemini CLI - For tasks requiring extensive context and sophisticated reasoning. Also excellent with multimodal and image understanding (for example, understanding screenshots of what you’re building).
- GitHub Copilot - Not very often anymore given the above tools
- Warp - AI-enhanced terminal that excels at complex command-line tasks and troubleshooting
API:
- I use APIs from OpenAI, Anthropic, and Google for programmatic access to their models
- OpenAI Realtime API for voice and streaming interactions
- MCP (model context protocol) servers for integrations
Chat-based interfaces:
- Claude (Anthropic) - Excellent for thoughtful reasoning, detailed analysis, and coding assistance, but I rarely use it via chat (versus via Claude Code) anymore
- ChatGPT (OpenAI) - Strong general-purpose assistant. I use the thinking models for complex problems.
- Google AI Studio (Gemini) - Exceptional for long-context tasks, multimodal input, and document analysis. See my Fever and Med Tracking project for an example of multimodal capabilities in action.
Mobile:
- Google Gemini app for multimodal live interactions (video and voice)
- ChatGPT app for advanced voice mode with vision capabilities
Coding 💻
Less and less in the IDE. Mainly powered by the tools mentioned above like Claude Code and Codex.
- VSCode (or other tools mentioned above) powered by the AI models mentioned earlier, but also as just another way to see the diffs for Claude Code or Gemini CLI
- Mostly Python and TypeScript. I’ve also used a variety of libraries (like Pandas, ScikitLearn, Numpy, and PyTorch).
Other tools 🔧
I’ve gotten great use out of all of the following for creating the presentation layer for projects. The last two I use for dynamic, interactive projects.
Personal note taking and management ✍️
- Dendron via VSCode