What's moving through AI, ranked by lead time.
Radar is a running index of what's moving through the AI ecosystem, ranked by where each story sits on the lead-time gradient: papers become repos become community threads become analysis. Every ingest run of the pipeline takes a fresh snapshot — several times a day, on no fixed clock — and the stamp below is the exact time of the one you're reading. The earlier a story is caught, the more head start you have before the crowd. Every row carries its receipts — which source saw it first, how many independent sources have it, how long ago — so the ranking explains itself. It's the public face of the same pipeline that feeds Vlad's Newsletter.
frozen 17 Nov 2026, 22:57 UTC · 67/78 feeds live · 16,065 tracked since 12 Jun
- 1 GPT‑6.2 Astra and Sol openai.com · 5h
- 2 [Research Note] Detecting Research Sabotage with Debate & Slopvestigators lesswrong.com · 18m
- 3 Show HN: R3 – Review AI agent artifacts locally r3.hyperlogue.tech · 42m
- 4 Show HN: I Cannot Let You In i-cannot-let-you-in.nodolabs.co.uk · 54m
- 5 It seems like misleading responses are mostly not intentional - Benchmarking LLM sophistry lesswrong.com · 1h
634 stories entered the watched feeds in the last 24 hours — 368 in papers, 7 in repos, 219 in community, 40 in analysis. The scored board skews to Community (40 of 40). 2 stories are heating across independent sources, and the median lead across the ledger's 14 observed fires is 14d — measured, not estimated. Freshest receipt, 10d ago: caught in GitHub Trending Python 119d before Show HN carried it. The oldest survivor has held the board for 5h.
- Repos → Community +119d lead GitHub Trending Python first · then Show HN 10d agovolcengine/OpenViking
- Papers → Community +22d lead arXiv cs.AI first · then Hacker News front 11d agoMemorilla: Latent Semantic Memory for LLMs
- Papers → Community +14h lead arXiv cs.AI first · then Hacker News front 16d agoProving and Improving the Platonic Representation Hypothesis with JEPAs
- 1GPT‑6.2 Astra and Solopenai.comLive Signal Community ·Models · 5h old · seen in 4 sources HEATING ×2
- 2[Research Note] Detecting Research Sabotage with Debate & Slopvestigatorslesswrong.comLive Signal Community · 18m old · via LessWrong
- 3Show HN: R3 – Review AI agent artifacts locallyr3.hyperlogue.techLive Signal Community ·Tools · 42m old · via Show HN
- 4Show HN: I Cannot Let You Ini-cannot-let-you-in.nodolabs.co.ukLive Signal Community · 54m old · via Show HN
- 5It seems like misleading responses are mostly not intentional - Benchmarking LLM sophistrylesswrong.comLive Signal Community · 1h old · via LessWrong
- 6How should we define “capabilities activities” for the purpose of pacing?lesswrong.comLive Signal Community · 1h old · via LessWrong
- 7Show HN: Even Keel – a task queue for coding agentsgithub.comLive Signal Community ·Tools · 1h old · via Show HN
- 8
- 9Show HN: I've asked AI to generate portraits per countryarnov.github.ioLive Signal Community ·Practice · 2h old · via Show HN
- 10Show HN: XControl – tmux groups for agents with dir sessions and peer messaginggithub.comLive Signal Community ·Tools · 3h old · via Show HN
- 11Show HN: A Background Agent to Fix Breaking Changes in Dependency Updatesarnaudrenaud.comLive Signal Community · 3h old · via Show HN
- 12Show HN: Whiz.pub – An AI-native blogging platform, free and donation-supportedwhiz.pubLive Signal Community ·Tools · 3h old · via Show HN
- 13Show HN: Open-source multi-player artifacts that update themselvesrallywith.aiLive Signal Community ·Tools · 3h old · via Show HN
- 14Show HN: Android system TTS engine backed by a self-hosted Qwen3-TTS servergithub.comLive Signal Community ·Tools · 3h old · via Show HN
- 15Show HN: Llx – Fast, Efficient GPT-6.2 Sol and Astra Coding via ChatGPTgithub.comLive Signal Community ·Tools · 4h old · via Show HN
- 16Show HN: Eso-box: turn a spare laptop into an always-on Claude Code servergithub.comLive Signal Community ·Tools · 4h old · via Show HN
- 17Launch HN: StarSling (YC P25) – Self-improving CI for agent-volume reposstarsling.devLive Signal Community ·Tools · 4h old · via Hacker News front
- 18Show HN: Fardesk – fast iPhone app to manage your Claude sessionsfardesk.franzai.comLive Signal Community ·Tools · 4h old · via Show HN
- 19Java4Shell: RCE in OpenJDK when processing Base64calif.ioLive Signal Community · 2h old · seen in 2 sources HEATING ×2
- 20Lorxus Does Budget Inkhaven For a Third Time: 11/11~11/17lesswrong.comLive Signal Community · 36m old · via LessWrong
- 21Quad9's withdrawal from France: A canary in the coal mineblog.cloudflare.comLive Signal Community · 42m old · via Hacker News front
- 22Show HN: Watch and listen to long-form x podcasts in one placexpodcasts.comLive Signal Community · 48m old · via Show HN
- 23Show HN: Grand Theft PTO – A browser game about skipping work for GTA 6labs.ficuscreative.comLive Signal Community · 1h old · via Show HN
- 24Show HN: Investing Won't Make You Richinvestingwontmakeyourich.comLive Signal Community · 1h old · via Show HN
- 25EK (emancipatory knowledge) in EA (effective altruism)lesswrong.comLive Signal Community · 1h old · via LessWrong
- 26
- 27Show HN: Code walkthroughs, maps, and quizzes for pull requestscodeunpacked.san.cxLive Signal Community · 1h old · via Show HN
- 28
- 29Show HN: I wrote a ProtectServer FM in Zig and skipped the vendor compiler808bits.comLive Signal Community · 2h old · via Show HN
- 30Show HN: Compact Contacts – a spreadsheet-style contacts app for Mac and iOScompactcontacts.comLive Signal Community · 3h old · via Show HN
- 31Show HN: Professor Euler – Offline math desk, 300-function CAS, workbookseuler.dawei7.comLive Signal Community · 3h old · via Show HN
- 32Show HN: Skif – An ultra fast alternative to Knip, Sherif and Syncpackgithub.comLive Signal Community · 3h old · via Show HN
- 33Show HN: A self-hosted, open-source database IDE in the browsergithub.comLive Signal Community · 3h old · via Show HN
- 34
- 35Web Browser Engineering (2021)browser.engineeringLive Signal Community · 3h old · via Hacker News front
- 36
- 37Effects of a Nuclear Detonation in Low Earth Orbitarxiv.orgLive Signal Community · 3h old · via Hacker News front
- 38350 years ago, Ole Rømer discovered that light has a finite speedsciencenorway.noLive Signal Community · 4h old · via Hacker News front
- 39Show HN: Isometric Montevideo a pixel art map built from Uruguay's open dataisometric-montevideo.pages.devLive Signal Community · 4h old · via Show HN
- 40Show HN: SSH into DOS: I wrote an x86 (IBM PC) emulatordialup.netLive Signal Community · 4h old · via Show HN
How Radar works
The gradient. Stories travel through the AI ecosystem in a rough sequence: they appear first in papers (arXiv, Hugging Face) or repos (GitHub), then surface in community discussion (Hacker News, Lobsters, LessWrong), then get interpreted in analysis (operator newsletters and essays). Radar ranks by where a story sits on that gradient. The earlier it is caught, the more lead time you have before it reaches consensus. The spectrum at the top of the page shows the last 24 hours of that flow — every story that entered the watched feeds, counted at the station where it first appeared.
The score. Each story gets a recency-decayed signal score. Every source tier carries a weight and a half-life, so live signal moves fast and fades fast while analysis sources move slowly and stay relevant longer. A story scores higher when two or more independent source families carry it — corroboration, not popularity, which is why five feeds of the same site count once. A small bonus applies when a story has visibly climbed the gradient. Stories past their tier's maximum age expire from the board.
Heating and lead time. A story is heating when it has surfaced in two or more independent source families. When the same story is observed first at an earlier stage and later at a higher one, Radar records the gap in hours — for example, seen on arXiv fourteen hours before it reached the Hacker News front page. That gap is the reader's lead time, and it is a measured span, never an estimate.
The classifier. Two judgments on this page are made by a model, not read from the feed log. A small typed classifier (TypeSafe's Jev) answers one question per story — is it about AI? — and stories it judges not to be about AI move to the bottom of the board; nothing is removed. It was checked against 232 hand-labelled titles before it shipped and agreed with 97.8% of them. The same classifier assigns the topic shown on a story and used by the filter above; a topic appears only when the classifier is confident, and at that setting it matched the labels on 115 of 116 stories. Stories without a confident topic are listed under All.
The honesty rule. Every other value on this page is an observed fact from the feed log — which source carried a story, when, and how many. Beyond those two labelled judgments, Radar makes no claim about what a story means; that is the newsletter's job. When a reading does not exist yet — no stories heating, nothing upstream — the page says so rather than inventing one.
- The spectrum band
- A filled flame tick is an early, lead-time-rich stage; a hollow tick is a later stage; a hollow ring on an empty span is a station with nothing on it yet.
- Signal class
- The tier chip: Live Signal moves fast and decays fast; analysis sources stay relevant longer. The ranker weights each accordingly.
- heating
- The same story in two or more independent source families — not two feeds of one site.
- seen in / via
- How many families carried it, or the single source that did.
Once a week, the operator read on what actually changed and the move it calls for. No digest, no hype.
Read the newsletter →