A recent analysis of Hacker News data reveals a striking shift: the percentage of Show HN submissions has climbed from a stable 2-3% over a decade to over 12% by late 2025. This surge correlates with the rise of LLMs like Claude Code and Cursor, which lower the barrier for non-coders to build and share projects.

While post volume grew, average scores tell a different story. Show HN stories, once matching general story scores, now trail by about 10 points. This suggests a community reaction to an influx of submissions, possibly from LLM-generated projects, though the analysis stops short of proving a direct quality link.

The author compiled this data using BigQuery and Python, sharing the code on GitHub. They note the difficulty in tracking LLM-generated content explicitly, as many posts don't disclose their tools. The trend indicates a fundamental shift in the platform's submission landscape, driven by accessible AI development tools.