Daily pipeline scores LLM coding model popularity and sentiment from top HN comments, mapped against the OpenRouter model catalog.
Key Takeaways
Pulls 200 top HN posts per day via API, filters to ~50 LLM/coding-relevant posts, then uses Gemini to identify models and rate per-comment sentiment.
Model detection is anchored to the OpenRouter catalog, making results auditable against a known model list.
All outputs logged to a Google Sheet with comment IDs, enabling direct audit by appending IDs to news.ycombinator.com/item?id=.
Dashboard shows a 10-day trailing aggregate of Top 10 model mentions and sentiment scores.
Hacker News Comment Review
Commenters note Claude leads in raw mentions but carries significant negative sentiment around API pricing and server downtime, while GPT-5.5 trails in mentions but has stronger positive feedback.
The choice of Gemini as the sentiment-rating LLM raised questions about undisclosed methodology and potential neutrality bias since Gemini is rarely discussed in coding contexts.
Basic UX issues flagged: model names on the chart are unreadable, limiting practical usefulness of the embedded visualization.
Notable Comments
@yakkomajuri: Asks which LLM drives the pipeline overall and notes Gemini’s selection for sentiment may reflect an attempt at neutrality given its low mention rate.
