AI has quietly slipped into every corner of modern development. Some use it as a copilot, others as a sparring partner, while a few continue to resist it altogether. Whether it’s generating tests, reviewing pull requests, or translating documentation, the tools have evolved, but so have the developers using them.

While full IDEs and code assistants have gained traction, a parallel story may be unfolding among those who prefer lighter, customisable environments. For these engineers, AI is not a replacement for deep work, but an accelerant that sharpens design thinking and preserves flow. It’s less about what AI can do and more about how it fits into an individual’s requirements.

Vim Over IDEs in the AI-Powered World

Dalibor Nasevic, a senior principal software engineer at GoDaddy, still uses Vim, but not out of stubbornness. He has integrated AI into a terminal-first workflow that now spans GitHub Copilot, Claude Code, Cursor, and local models via Ollama and LM Studio. 

For context, Vim is an open-source text editor that was first released in 1991 and remains in the top five list among developers due to its speed, efficiency, and ubiquity across Unix-based systems. Unlike modern editors, Vim is entirely keyboard-driven, enabling power users to navigate and edit code with minimal keystrokes. 

It runs in the terminal, making it lightweight and accessible on remote servers without a graphical user interface (GUI). Though it has a steep learning curve, Vim is endlessly customisable through configuration files and plugins. For many, Vim offers unmatched precision and control, which is why it has continued to thrive decades after its release.

There are various developers, including Nasevic himself, who prefer Vim over VS Code and others.

“It works well in the terminal and doesn’t enforce a built-in editor, which is ideal since I prefer working in Vim rather than switching to a full IDE.”

At GoDaddy, experimentation with AI tools is actively encouraged. For Nasevic, however, it’s not about flashy output. “As long as I understand what’s going on under the hood…I find AI incredibly helpful,” he writes in a blog post.

His AI journey began with casual queries and second opinions. It soon grew to encompass language conversions, API specification writing, and full-task delegation wherever appropriate. Claude Code has become his favourite, especially when used through Amazon Bedrock, which allows for better security and privacy while keeping him inside his preferred editor.

Yet, Nasevic is clear-eyed about where AI fits. Tools can hallucinate, especially when data is patchy or prompts are vague. “I’ve seen some models hallucinate or loop on vague prompts, especially when trained on incomplete data,” he wrote.

Whether it’s generating flawed AWS security rules or misinterpreting new product APIs, he’s learned that accuracy often hinges on the tool’s ability to set context, not just on the model itself.

AI is Helping, But One Should Be Cautious

Rather than interrupting his work, AI has helped Nasevic preserve momentum. He is able to move seamlessly between design, code, and infrastructure without friction. “It doesn’t make decisions for me,” he explained. “It simply sharpens my thinking and speeds up the process.”

The developer role is evolving. Nasevic sees it tilting toward orchestration, where judgment, clarity, and contextual understanding matter more than typing speed. With the help of AI, engineers can stay in control while reducing cognitive load. However, vigilance remains essential. 

“While it can boost productivity, it can also create false confidence, which may become risky and costly,” he wrote.

And through it all, Vim remains in the picture. 

AI Doesn’t Change The Developer Spirit

Programming is still a long game. Nasevic believes short-term gains should not come at the cost of long-term clarity. “Find what you enjoy,” he further wrote. “Do the work that excites you.” In his case, it’s building systems, integrating smart tools, and never quite letting go of Vim.

Nasevic mentioned that distinguishing oneself requires foresight. Currently, this entails integrating AI and discovering adaptive strategies that preserve your advantage.

The post A GoDaddy Engineer’s Take on Using AI Without Losing Vim appeared first on Analytics India Magazine.