This is a fairly good selection of capabilities, and I'm finding myself using it in similar ways. Aside from a bunch of the tools in this article like Aider and Cursor I'm also using Dify to ingest my Obsidian notes so that I can ask questions and I've wired this up to lobechat for a generic interface that also allows me to use a broad range of models.
I've got a 96GB M2 Max Macbook so I'm using a lot of local models where possible. The dify setup is all local ollama for both embedding (mxbai-embed-large) and llm (llama3:70b,command-r,gemma2)
I've also been using superwhisper to do a lot of voice typing.
It's not a simple setup. You have to use something like https://github.com/kim-space/dify2openai because dify doesn't implement the openai endpoints. There may well be better options out there now, I've been running this setup for a while.
Even though Anthropic is more focused on providing a service to other companies, I hope they will also provide a voice mode to communicate with their models, I prefer using Claude 3.5 Sonnet compared to GPT-4o.
I've got a 96GB M2 Max Macbook so I'm using a lot of local models where possible. The dify setup is all local ollama for both embedding (mxbai-embed-large) and llm (llama3:70b,command-r,gemma2)
I've also been using superwhisper to do a lot of voice typing.