Show HN: What is HN thinking? Real-time sentiment and concept analysis

Source: Hacker News

Overview

Hi HN, I made Ethos, an open‑source tool to visualize the discourse on Hacker News. It extracts entities, tracks sentiment, and groups discussions by concept.
Check it out: https://ethos.devrupt.io

This was a “budget build” experiment—I managed to ship it for under $1 in infrastructure costs.

Technical details

• Initially used qwen3-8b for the LLM and qwen3-embedding-8b for embeddings.
• Ran into capacity issues with that model, so switched to llama-3.1-8b-instruct to stay within a similar budget while achieving higher throughput.

Questions:

• What LLM or embedding would you have used within the same price range? It needs to support structured output.
• How problematic is it that llama-3.1 is used with a higher‑dimensional embedding? I originally wanted the LLM and embedding to be from the same family, but I’m not sure if there’s a “munch point” in that.

Repository

GitHub – devrupt-io/ethos

Feedback

I’m looking for feedback on which metrics (sentiment vs. concepts) you find most interesting! PRs welcome.

Discussion

Comments on Hacker News (7 points)