Instead of Claude and Anthropic models, I use my local LLMs for coding (and not vibe-coding, mind you)
…Heck, even if they don’t use it to train new AI tools, I can’t risk submitting even a snippet of the proprietary source code from classified projects. Related I automated…