HN reader with sentiment filtering of comment threads
Problem
HN readers feel comment sections have become overwhelmingly pessimistic and cynical, especially on AI-related launches, and manually collapsing negative threads on every post is exhausting. Existing aggregators offer no way to control the tone of what you read.
Opportunity
An HN reader/aggregator with per-thread sentiment classification, letting users filter or deprioritize cynical/negative comment trees and surface constructive or celebratory discussion first.
Market analysis
The pieces already exist scattered around: HN-Clean collapses negative comments via user-defined thresholds, Hackermoods runs per-story sentiment analysis, and Comments Owl handles thread filtering and muting. Gluing them into a full reader with per-thread sentiment classification is a weekend project — which is also the problem, because there is no defensible layer and no evidence anyone would pay.
Market · HN power readers burned out on cynical AI threads; relatable pain, but a small audience that expects free tools and is itself skeptical of the premise.
Pricing · Effectively zero willingness to pay: HN readers and extensions in this space are free, and the audience is famously monetization-hostile — at best a free extension with a tip jar.
Pros
- + Cheap to build on the public HN API; classification can run client-side.
- + Sentiment tooling on HN (Hackermoods) already proves the pipeline is feasible.
- + Per-thread tone control is a genuinely novel twist over story-level filtering.
Cons
- − HN-Clean already collapses negative comments by threshold; Comments Owl covers filtering and muting.
- − No credible revenue path: every comparable HN reader or extension is free.
- − Classifiers struggle with HN-grade sarcasm and will mislabel valid technical criticism.
- − The product's core premise — hiding criticism of AI launches — invites the very backlash it tries to filter.
Existing / similar tools
Source
Hacker News (Ask HN)
Treat this as open-source portfolio material, not a product. The technically interesting problem hidden inside it is tone-aware ranking that survives HN sarcasm — naive sentiment models will score a dry, valuable technical takedown as negative and a sarcastic cheer as positive, and users will notice within a day. If anything is worth building, it is a classifier calibrated specifically on HN comment style, released as a library others can bolt onto their readers; owning that component is more credible than owning yet another HN client.