Various AI tools that I am using which will change before I finish writing this.
AI Tools
- Herdr
- OpenRouter
- Codex
- ChatGPT
- Buzz
- OpenCode
- T3Code
- ClaudeCode
- Linear
- OpenAI Symphony
AI Models
- GPT 6 Astra
- Claude Fable 5.1
- GLM 5.2
- QWEN 3.8
- Gemma 4 for local
- Eleven Labs
- GPT Image 2
Skills
Possible Future Investigations
Unsloth
Unsloth is an open-source framework for running and fine-tuning models locally with less memory and faster training. I want to try it for adapting open models such as Gemma and Qwen on my own hardware.
Diagram Design
Diagram Design experiments with AI-native tools for user interface and product design. I am interested in seeing whether it can speed up the process of turning rough ideas into useful visual designs.
omp
omp is an open-source terminal coding agent with built-in IDE features, browser control, debugging, and subagents. I want to see how its extensible, multi-model harness compares with Codex and OpenCode.
bb
bb is an open-source, local-first IDE for working with coding agents such as Claude Code and OpenCode. Agents can run in separate threads and customize the workspace through plugins. I want to see whether bringing agents into one workspace makes it easier to follow their work and switch between tasks.
GitNexus
GitNexus from YC https://www.akonlabs.com reduces the number of tokens sent to models by creating a knowledge graph of the source code. Unfortunately I can see this becoming part of the agent harnesses as they look for ways to improve their tools. I expect one of the bigger players will buy the company and add it to their product.