Was tired of drowning in HN comments, so I built an AI Chief of Staff
I've been lurking on HN for years. You know the drill: interesting headline, 200+ comments, you dive in thinking "I'll just skim for 5 minutes"... and an hour later you're 36 chambers deep in a thread about memory allocation patterns in Postgres and you've completely forgotten what the original article was about. I don't just want a "summary" (which usually just shortens the noise). I want the meta-consensus: "What is the actual trade-off being debated? Who is winning the argument? Why does this matter?" So I built HNSignals. Think of it less like a "summarizer" and more like a Chief of…
What it does
In the maker’s words, at launch
I've been lurking on HN for years. You know the drill: interesting headline, 200+ comments, you dive in thinking "I'll just skim for 5 minutes"... and an hour later you're 36 chambers deep in a thread about memory allocation patterns in Postgres and you've completely forgotten what the original article was about. I don't just want a "summary" (which usually just shortens the noise). I want the meta-consensus: "What is the actual trade-off being debated? Who is winning the argument? Why does this matter?" So I built HNSignals. Think of it less like a "summarizer" and more like a Chief of Staff who reads the entire thread for you and hands you a one-page executive brief. How it works: 1. Filters: Shows trending stories (50+ comments) where the discussion has heated up. 2. Extracts: An AI (Qwen 3 via Venice.ai) reads the top comments. 3. Structures: Instead of a wall of text, you get 4 specific signals: - The Hook: Why you should care. - The Gist: The core technical facts. - The Debate: The actual friction point (e.g., "Rust vs. C++ safety"). - The Verdict: The community consensus. Why I'm showing this (Beta): Most AI tools just shorten the text. I'm trying to extract the signal. It's a work in progress - the AI sometimes gets too clever, and I'm still tuning the cache strategy. I'd love feedback on whether this structured approach is actually better than a standard "TL;DR." Bonus meta-game: If this Show HN gets 50+ comments and makes it onto HNSignals itself, I'll read the AI's analysis of people analyzing my analyzer. (P.S. Please go easy on the intentional stress testing - my Lambda inference budget is finite!) Try it: https://hnsignals.com Example output The Heartbleed Bug (2014): https://hnsignals.com/signal/7548991 Ask HN: What is the most unethical thing you've done as a programmer? (2018): https://hnsignals.com/signal/17692005 Comparing the Same Project in Rust, Haskell, C++, Python, Scala and OCaml (2019): https://hnsignals.com/signal/20192645 OpenAI's GPT-3 may be the biggest thing since Bitcoin (2020): https://hnsignals.com/signal/23885684 Apple, What Have You Done? (2026): https://hnsignals.com/signal/46763592
Does the same job
all alternatives →- ASA self-hosted AI that turns Hacker News into a daily briefingJul 2026 · github.com · ▲5
Hi everyone, I read HN every day, but there was always more great content than I had time to read. I know there are already several HN summarizers on GitHub, and I tried some of them. They just didn't fit the workflow I wanted, so I decided to build my own. My project is a self-hosted app that automatically fetches top stories, summarizes them with AI, translates them into your preferred language, and prepares a personalized daily briefing in a nice customizable UI while you sleep. Maybe I'm not the only one who wanted this kind of workflow. I'd love to hear your thoughts, especially on what…
- IMI measured the half-life of 41,301 Show HN launches. It's 7 hoursJul 2026 · jonno.nz · ▲37
I scraped every Show HN from the last 12 months (41,301 posts) plus the full comment tree of every launch with 10+ comments, ~100k comment timestamps, all from the Algolia HN API. The median launch gets 2 points and 0 comments. For launches that do get traction, half the comments they'll ever get arrive within 7.2 hours and 90% within 26, and the top decile decays on the same clock as everyone else. Vote timestamps aren't public, so comment timing is the attention proxy; caveats are in the post. Everything reproduces from the repo with one command…
- HMHubski, my idea of a better aggregator2011 · ▲18
One year ago, I downloaded the news.arc code and decided that I would teach myself programming by building a better aggregator. At that time, I had some rudimentary HTML knowledge, and little FORTRAN under my belt. I knew that I wanted an aggregator that could support a wide range of submissions, and one that could scale without losing quality for the user. Hubski is what I have to show for my time and effort. The easiest way to describe Hubski is to say that it uses a commenting system akin to HN or Reddit, but uses a sharing scheme like Twitter. You have a personal feed comprised of the…
- BEBrowser extension to summarize HN comments – bring your own AI models2024 · github.com · ▲8
We’re George and Ann, and want to share a Hacker News specific browser extension that we have been working on. We all love the rich discussions in HN, but navigating long posts with multiple threads can sometimes be challenging. You get lost in the depth and breadth of the discussion. There are a few AI-based summarization tools, but they typically focus on summarizing the article rather than the comments, and thus miss the essence of threaded discussions. So, we built Hacker News Companion, a browser extension that makes it easy to navigate discussions and understand diverse viewpoints…
- KTKeep track of what different AI's say about you2023 · haveibeenencoded.com · ▲5
Hi HN! I built this tool, because Large Language Models are hallucinating their asses off and I wanted to test just how bad it is with a topic I know best - myself. I'm sure there are other egos out there who google themselves and essentially this is the new googling yourself. It's early beta, so lots of room for improvement of course.
- WBWe built an AI Website builder with better output2024 · dorik.com · ▲8
Hey HN, After GPT-3 created waves in the tech industry, a lot of AI tools were emerging and with that, some AI website builders But the results seemed way too generic to us. It felt like the developers were rushing to catch the wave instead of building a proper tool We took our time, did months of RnD and finally came up with something better than what others in the market are doing. It’s got better design output. While it’s still in beta, I wanted to show HN what we did. Will appreciate the feedback when you guys try it out. Here is the link to signup for the beta:…
More ai this month
the category →
I trained a 125M-parameter transformer to autocomplete piano performances in real time (~108 notes/sec on an iPhone 15). The idea is basically GitHub Copilot or Tabnine, except instead of prompting it with code, you prompt it by playing a few notes on a MIDI piano. The model then continues what you played, entirely on-device. The app is free if anyone wants to try it. Happy to answer questions about the model, training, Core ML, or the many things that didn't work.
AI · 17d ago · simedw.com
Astute▲585Automate your B2B brand going viral, with new media creators
AI · 18d ago · company-app.joinastute.com
Hey HN, Henry from Cactus here! We previously released Cactus Needle, a 14MB agentic LLM for tool call, device use, and structured extraction for phones, wearables, smart homes, small robots and microcontrollers. We got really great feedback here, and have now incorporated the suggestions to release Needle 2. The whole model is a single 14MB binary that runs a full session in 28MB of RAM; 45m parameters at 2bit compression. Needle hits 500 tokens/sec decode speed on a Raspberry Pi 5, sits between 400-1,500 tokens/sec on VR devices like Meta Quest 3S and Apple Vision Pro, and ranges…
AI · 27d ago · cactuscompute.com
Launched alongside, January 2026
the whole month →- IN
Hey HN! I wanted to share something I built over the last few weeks: isometric.nyc is a massive isometric pixel art map of NYC, built with nano banana and coding agents. I didn't write a single line of code. Of course no-code doesn't mean no-engineering. This project took a lot more manual labor than I'd hoped! I wrote a deep dive on the workflow and some thoughts about the future of AI coding and creativity: http://cannoneyed.com/projects/isometric-nyc
AI · Jan 2026 · cannoneyed.com
Automatic AI-powered code reviews the moment you open a PR
Dev tools · Jan 2026 · kilo.ai
