For the past two years I’ve worked on a cybersecurity platform for enterprises. They way I work is that I use GNU Nano to write code (Python). I consult various AIs, but I don’t allow an AI to touch my code files, HW or network. All code is entered manually by myself, and I research any advice I’m given my a coding buddy or a AI coding buddy. I do understand that IDEs give me certain benefits, particularly when you work in teams and have a large code base. I don’t participate in AI agents or vibe coding.

The obvious reason I work like this is to reduce the attack surface. I also like having ownership and learn more as I build new features. Deep code understanding and error awareness are the main benefits. But, I could be more productive and less error prone with the right tools. Also, security audits demand traceability.

I readily admit that manually typing and entering code seems quaint when you work in an industry that’s racing towards automation. So my question to you is what analytical, type checking, dependency scanners, and integrations with terminal would be beneficial for my productivity?

What is your assessment of the process?