This Hacker News discussion centers around using Large Language Models (LLMs) like ChatGPT for coding. The original post, a Medium article titled "Vibe coding is a hobby," sparked debate about LLMs’ role in software development. Several commenters report *positive* experiences. One user leverages LLMs for tedious tasks, focusing their energy on problem-solving and then refining the LLM’s output. Another claims to consistently produce high-quality code and faster development cycles *at work* using LLMs under their direct control. A contrasting viewpoint links the discussion to a recent statement suggesting the end of human-written code, but a reply clarifies the article advocates for approaching LLMs with a learning mindset – similar to acquiring a new language – for optimal results. Ultimately, the conversation highlights a shift towards using LLMs as tools to *augment*, rather than replace, human coding skills.