As AI gets more expensive, popularity of Chinese AI models like DeepSeek is growing fast, even though these models are not as good as ChatGPT, Claude and Gemini. Industry experts believe the cost pressure is real and most companies will move to a hybrid model for AI use.
Does anyone use deepseek for programming, what does your workflow look like?
The easiest way it to use opencode, you get an API key from DeepSeek and put it in your shell environment. Then fire up opencode in the project folder and ask the model to give you a phased plan for the feature you want it to implement, review it, then tell it to build it one deliverable at a time, on a branch, review the work at each step, and go to the next. Doing things in phased fashion is the easiest way to make sure things stay on track and you understand what the code is actually doing.
I don’t use deepseek, but I do use qwen3.
I mostly use this: https://github.com/madox2/vim-ai
I open a file, and ask AI to write something. Then I start adjusting it, or fixing the mistakes.