Show HN Posts Are Rising While Their Scores Are Falling

{{IMAGE:1}}

I've been noticing more "Show HN" posts on Hacker News lately, and many seem to be built with AI assistance. So I pulled the data to check if my perception matched reality. The numbers tell a clear story: Show HN submissions are skyrocketing, but the community's reception is cooling.

The Trend: A Decade of Stability, Then a Surge

For nearly ten years, from 2012 to 2022, Show HN posts made up a consistent 2-3% of all stories on Hacker News. The pattern was remarkably stable. Then, around 2023, something shifted. The percentage began climbing steadily.

The acceleration is unmistakable. Tools like Claude Code and Cursor 1.0, which allow developers to generate functional code with natural language prompts, have made it easier than ever to create something worth sharing. By December 2025, Show HN posts accounted for over 12% of all stories—a fivefold increase from the historical baseline.

This isn't just a minor fluctuation. It represents a fundamental change in who can participate in the "show and tell" corner of the developer community. The barrier to entry for creating a shareable project has lowered dramatically.

The Score Decline: Community Reception Wanes

What makes this trend particularly interesting is how it correlates with community reception. Historically, Show HN posts received scores comparable to other stories—typically between 15 and 18 points. That parity held through 2023 and into 2024.

Then the divergence began. As the percentage of Show HN posts increased, their average scores started falling. By December 2025, the average Show HN score had dropped to 9.04 points, while the average for all stories remained at 19.53—a gap of over 10 points.

This creates a fascinating tension. On one hand, more people are creating and sharing projects. On the other, the community appears to be growing less enthusiastic about these submissions.

What's Driving the Change?

Several factors could explain this pattern:

1. Volume vs. Quality The sheer increase in submissions might be diluting individual scores. When fewer than 3% of posts were Show HNs, each one felt special. Now at 12%, they're a common sight, and the novelty has worn off.

2. Perceived Authenticity Many Show HN posts are now built with AI assistance. While this doesn't inherently make them lower quality, the community might be developing a bias against "AI-generated" projects. There's an implicit question: "Did the creator actually build this, or did they just prompt an AI?"

3. The 2022 Anomaly Interestingly, average scores for Show HN posts actually increased in 2022, bucking the overall trend. This could indicate a temporary surge in high-quality submissions or perhaps a shift in community voting patterns. Without deeper analysis, it's hard to pinpoint the exact cause.

The Data Challenge: Identifying AI-Generated Content

One limitation of this analysis is that we can't definitively identify which Show HN posts were built with AI. Many creators don't mention using LLMs in their descriptions, and there's no reliable way to detect AI-generated code from the metadata alone.

This creates a blind spot in the analysis. If AI-generated projects are indeed receiving lower scores, we can't confirm whether this is due to their quality, their origin, or simply the increased volume.

The Broader Pattern

This trend reflects a larger shift in software development. Tools like Cursor and Claude Code are democratizing creation. People who never learned traditional programming can now build functional applications. This is both empowering and disruptive.

The Hacker News community's response suggests a complex reaction. They're seeing more projects, but their enthusiasm is waning. This could be a temporary adjustment period, or it might signal a lasting change in how the community values different types of contributions.

The Data and Code

For those interested in the methodology, I've shared the Python code and CSV data on GitHub. The analysis uses BigQuery's public Hacker News dataset, filtered for story-type posts with titles.

The challenge of identifying Show HN posts from the BigQuery data is that the type field doesn't include a "show_hn" attribute. I worked around this by lowercasing titles and filtering for those starting with "show_hn: ".

The original CSV was too large (~400 MB) to commit directly, but it's freely available through BigQuery. I ran the query, exported to Google Drive, and downloaded it for local analysis.

Future Observations

This analysis is just a snapshot. The trend is clear, but the underlying reasons require more investigation. I plan to update this analysis periodically to track whether the score decline continues or if the community finds a new equilibrium.

The key question remains: Is the community reacting to the quantity of Show HN posts, their perceived quality, or the changing nature of what it means to "build" something worth sharing? The data shows the what, but the why will take more observation.

For now, the numbers speak clearly: Show HN is experiencing unprecedented growth, but with diminishing returns in community appreciation.