Vibe Coding Is About a Year Old

TL;DR AI
2 min readKey summary
A year after Andrej Karpathy coined “vibe coding,” the conversation is shifting toward “agentic engineering.”
The new framing emphasizes human oversight, software quality, and keeping developers in control of AI-generated code.
Stanford research suggests AI coding tools can make programmers feel more confident while producing less secure code.
Products and discussions around Lovable, Blitzy, and enterprise use cases show AI-assisted coding moving beyond novelty into broader workflows.
The big tradeoff: faster code generation can boost productivity, but weak oversight raises security and debugging risks.



