- Author
- Posted by e12e, with replies from many commenters
- Published
- 6 May 2026
- Kind
- Hacker News discussion thread on a blog post
- Original
- news.ycombinator.com/item?id=48037128
- Subject
- Coding with agents
What it is
The linked post is by Simon Willison, dated 6 May 2026. It records highlights from a podcast about AI coding tools. The thread gathers replies from many commenters.
What it says
- Willison's post says vibe coding and agentic engineering have started to blur for the author.
- The post says the author no longer reviews every agent-written line, and feels guilty.
- The post says the author values proof of real use over tests or docs.
- The post says the software lifecycle assumed a few hundred lines of code a day.
- Commenter singpolyma3 says it matters little whether a human or an LLM wrote code.
- Commenter jf22 replies that most software was already a mess.
- Commenter yakattak says people who ignored quality can now make more mess faster.
What evidence it gives
The post is the author's reflection from a podcast, with no data. The thread is opinion and anecdote, with some analogies and predictions. Few comments cite sources.
Interests and limits
The commenters are self-selected. Predictions about code in 30 years are speculation. Read the linked post in full but only the opening comments of the thread.