AI writes 97% of our code. Strong engineers still matter
At MedWrite, AI generates 97.1% of committed code, allowing engineers to focus on reviewing and refining outputs rather than writing code from scratch. Despite heavy automation, strong engineering skills remain critical for catching context-specific errors, ensuring compliance, and maintaining system security. The team rejects more AI-generated code than it accepts, underscoring the need for deep technical judgment. AI accelerates development but does not replace the need for experienced engineers in high-stakes domains like healthcare.
Opening excerpt (first ~120 words) tap to expand
We Automated 97% of Our Code with AI. Here’s What We LearnedWhy coding skills still matter more than everAhmad Albarqawi6 min read·Apr 17, 2026--ListenSharePress enter or click to view image in full sizeAI operator engineerAt my startup MedWrite, we started using AI for software development early in 2025. We did not adopt it because it was trendy. We adopted it because we are a startup competing in healthcare, and we wanted a practical way to move faster than companies with much larger engineering teams. Our backlog was growing quickly from doctor feedback and insights from the field and we wanted to increase each engineer’s output safely.
…
Excerpt limited to ~120 words for fair-use compliance. The full article is at Medium.
Discussion
0 commentsMore from Medium
