The debate over AI coding tools has intensified as developers split between those embracing automation and those skeptical of outsourcing software engineering. A recent LinkedIn post claiming top engineers now use AI for coding sparked heated discussion, with critics arguing these tools can't build reliable software. The reality is more nuanced - while AI coding tools have advanced rapidly, they're changing rather than replacing engineering work.

Cursor launched in 2023 but struggled with reliability until recently, while Anthropic's Claude Code has surged in popularity. Spotify claims its top developers haven't written manual code since December, and Anthropic reports 80% of deployed code comes from AI assistance. Andrej Karpathy noted programming changed more in the last two months than in years prior. These tools still require human oversight - planning, architecture, and system thinking remain critical. Engineers must direct projects, research tech stacks, and understand tradeoffs before delegating to AI.

The shift toward AI-assisted development demands new skills. Success requires judgment about what to question, when to push back, and how to evaluate quality. This "taste" for good architecture and UX can be learned rather than innate. While tools like Claude Code dominate current conversations, the fundamental engineering work of designing systems and making architectural decisions persists. Software engineering isn't disappearing - it's evolving into a role where humans orchestrate AI agents rather than writing every line manually.