Developer summarizes a year of experience with AI coding
Understand this story
A developer summarizes observations from a year of using AI to write code: code is no longer "written" line by line, but rather "grows" like seeds, with each project first having a "constitution"-style document explaining goals and design principles. He argues that codebases are no longer read directly by humans, but retrieved and distilled by AI agents, so the README needs to clearly explain purpose, structure, and usage. He also suggests having AI do only one thing at a time and maintaining strict structural rules to prevent AI from repeatedly rewriting the same function. These are the author's personal experience summaries, not proven results.
AI overview based on source reports · Updated 22 hr ago
Coverage timeline
Follow the coverage to explore different perspectives.
- Hacker News AIThings I learned coding with AI for a year
A developer reflects on a year of AI-assisted coding, describing how code is now "grown, not created" from constitution-style documents, that codebases are inspected by proxy through AI rather than read directly, and that systems are refined via rapid rewrites and alternative-design explorations. The post details practices such as metaplans for context, domain/layer structural rules, cleanup loops to combat "slop," and the view that design matters more than ever while some model behaviors can't be fought.
Activity for this story
- Current comparable activity
- 5
- Peak comparable activity
- 10Nov 21 10:00
- Change in the past 24 hours
- –
Trends compare the same fully observed participants, which may cover fewer participants than current activity. Move the pointer or tap the chart to inspect hourly activity; use the left and right arrow keys with a keyboard.