Morning Brief: Saturday, November 21

Seventy-one feeds reporting. Two weeks. 5,515 items reduced to what follows. (what we track, how we crawl, subscribe)

Saturday's thread is proof. The Riemann Hypothesis was claimed twice this week by different parties using different methods – one of them machine-assisted – in the same days that formal verification was pitched as the only durable answer to the volume of machine-written code.

The two Riemann claims are not a single story. Levent Alpöge posted a remark.pdf under his own name; separately, Slashdot reports an Anthropic mathematician claiming a proof, and TechCrunch frames the pair as an AI claim and a human claim arriving together. Scott Aaronson's post title – The stochastic parrots have apparently proved the Riemann Hypothesis – registers the obvious tension without settling it, and Terence Tao published on the Generalized Riemann Hypothesis the same day in a different register entirely (unforced Euler and Navier-Stokes singularities). Simon Willison quoted both Ravi Vakil and Alpöge. Nothing here is a refereed result; what is notable is that the referee question now has two different provenance stories to resolve rather than one.

Anthropic's IPO round-tripped inside the same week. It priced at $83 and opened at $111; the 2026-11-19 reporting has it topping a $2T market cap on debut, Anthropic announced the offering's close with the underwriters' option fully exercised on the 20th, and the 2026-11-20 reporting has shares tanking as the Nasdaq erased its 2026 gains. The seven founders took out nearly $2B; the structure could route $156B to charity. WBUR's Here & Now ran the Wall Street read on the 20th.

Top (5-7 min)

Levent Alpöge Claims Proof of Riemann Hypothesis
Lobsters, 2026-11-20. The human claim, posted as a PDF under the author's own name. Read this before reading anyone's commentary about it.
The stochastic parrots have apparently proved the Riemann Hypothesis
Shtetl-Optimized, 2026-11-20. Aaronson on the machine-assisted claim. The title is doing the skeptical work; the post is the closest thing to a complexity theorist's triage note on the week.
AI claimed a solution to the 167-year-old Riemann hypothesis – and a human did, too
TechCrunch, 2026-11-20. The framing that makes this a provenance story rather than a math story: two claims, two methods, one referee problem.
Unforced Euler singularity, unforced Navier-Stokes singularity, and the Generalized Riemann Hypothesis
Terence Tao, 2026-11-20. Published the same day and unrelated to the claims above. Included because it is what working on this family of problems looks like when nobody is announcing anything.
The War Against Slop and an Emerging Weapon to Fight it: Formal Verification
Latent Space, 2026-11-20. The explicit argument that verification is the answer to volume. Pairs with the Anthropic code-review piece below and with yesterday's Pandora's Toolbox negative result.
Anthropic says Claude writes and reviews 90% of its team's code. The catch is how it checks the work.
The New Stack, 2026-11-20. A number worth less than the mechanism behind it. The reported catch is the checking layer, which is the same claim Latent Space is making in the abstract.
AI R&D-relevant model capabilities are improving rapidly and are getting harder to measure
METR, 2026-11-20. METR's Q4 update says both halves out loud. The measurement half is the one that matters for every benchmark in the scan below.
Anthropic Shares Tank After Huge IPO As the Nasdaq Erases 2026 Gains
Slashdot, 2026-11-20. Two sessions after the $2T print. The index move and the single-name move are not separable from this reporting.

Themes

Two claims, two provenance stories

The interesting property of this week is not that the Riemann Hypothesis was claimed. It is that it was claimed twice, by routes that require different kinds of checking. A human-authored PDF gets read by referees who can interrogate the author. A machine-assisted claim raises the question of what exactly is being attested and by whom – which is the same question the verification papers in the scan are asking about agent output in general, at much lower stakes.

Levent Alpöge Claims Proof of Riemann Hypothesis
Lobsters, 2026-11-20.
Anthropic Mathematician Claims to Prove Riemann Hypothesis
Slashdot, 2026-11-21.
[AINews] Riemann Hypothesis proved…twice independently
Latent Space, 2026-11-21.
Quoting Levent Alpöge
Simon Willison, 2026-11-20.
Quoting Ravi Vakil
Simon Willison, 2026-11-20.
A Tensor-Algebraic No-Go Theorem for High-Dimensional Photonic GHZ States Discovered with AlphaProof Nexus
arXiv cs.AI, 2026-11-20. A machine-discovered no-go theorem in a domain where the result is checkable. The contrast with an unrefereed Riemann claim is the whole point.

Verification as a response to volume

Three unrelated sources argued the same structural thing this week: when the volume of generated artifacts rises, the binding constraint moves from production to checking. Latent Space makes it as a thesis, Anthropic makes it as a practice, and Antithesis and the ACM paper below make it as engineering that already shipped against Raft and against a production object store.

The War Against Slop and an Emerging Weapon to Fight it: Formal Verification
Latent Space, 2026-11-20.
Verifying formal models of Raft implementations
Antithesis, 2026-11-19. Formal models checked against implementations rather than admired separately from them.
Validating a High-Performance Cloud Object Store with Lightweight Formal Methods
ACM DL via Lobsters, 2026-11-19. The word doing the work is "lightweight". This is the tractable end of the formal-methods spectrum, applied to something in production.
Verify the Verifier: Specification-Grounded LLM Cross-Checking for Bugs That Escape UVM Scoreboards
arXiv cs.AI, 2026-11-20. The regress made explicit: who checks the checker, in a domain (hardware verification) that has had scoreboards for decades.
Introducing Frostbite: A new CPU architecture discovered, designed, and verified end-to-end by AI
Architect Labs, 2026-11-21. Vendor announcement, read it as such. Worth logging because "verified end-to-end" is the claim being made, and the arXiv EDA cluster below is the research-side version of the same bet.

The IPO round trip

Anthropic Tops $2 Trillion Market Cap After IPO Debut
Slashdot, 2026-11-19.
Anthropic hits Nasdaq exchanges at $111, 33% above IPO price
TechCrunch, 2026-11-19. The source for the $83 pricing and the $111 open. Note the URL slug says $110 and 27%; the headline says $111 and 33%. Published as the feed gave it.
Anthropic announces closing of initial public offering, including full exercise of underwriters' option
Anthropic, 2026-11-20. The primary document for the mechanics: offering closed, greenshoe fully exercised, concurrent private placement closed.
Anthropic Shares Tank After Huge IPO As the Nasdaq Erases 2026 Gains
Slashdot, 2026-11-20.
Anthropic's IPO puts AI's circular financing boom to the test
TechCrunch, 2026-11-19. The structural question the price action is now answering in public.
Anthropic's 7 founders cashed out nearly $2B in record-breaking IPO
TechCrunch, 2026-11-19.
Anthropic's IPO could deliver $156B to charity
TechCrunch, 2026-11-19.
At Nasdaq opening bell, Anthropic CEO recommits to pacing AI development
TechCrunch, 2026-11-19.
Analyzing Anthropic's Wall Street debut
WBUR, 2026-11-20.

Scan

Chip design gets its own verification agents

Eight papers in one arXiv announcement treat RTL, UVM and EDA as agent-harness problems. This is the most concentrated domain cluster in the window, and it is notable that the domain already had formal verification culture before the agents arrived.

Verify the Verifier: Specification-Grounded LLM Cross-Checking for Bugs That Escape UVM Scoreboards
arXiv cs.AI, 2026-11-20.
DVBench: End-to-End Benchmark for Real-World UVM Design Verification Agents
arXiv cs.AI, 2026-11-20.
SiliconGym: A Versatile Reinforcement Learning Environment for Agentic EDA
arXiv cs.AI, 2026-11-20.
Meta-RTL: Evolving Agent Harnesses for RTL Optimization
arXiv cs.AI, 2026-11-20.
GRODE-RTL: Source-Mapped Dependency Views for LLM-Based RTL Repair
arXiv cs.AI, 2026-11-20.
FPGA-GEN: Target-Aware RTL Generation via HLS
arXiv cs.AI, 2026-11-20.
From FIPS to FPGA with LLM Agents: Measuring What Tests and Reviews Miss in Post-Quantum Cryptographic Hardware
arXiv cs.AI, 2026-11-20. The measurement framing – what tests and reviews miss – is the useful one, and post-quantum crypto hardware is a good place to ask.
CHORUS: Coverage-Guided Agentic RL and Policy Consolidation for Testbench Stimulus Generation
arXiv cs.AI, 2026-11-20.

Agents, memory, and harnesses

Your AI agent's "memory problem" may be something else
The New Stack, 2026-11-20.
Does your AI agent need an LLM to decide what to do next?
The New Stack, 2026-11-20. AWS decider models: the routing step pulled out of the model.
A "headless Slack": Vercel's new message board lets Claude Code and Codex agents work together
The New Stack, 2026-11-19.
An engineer didn't message his AI for 22 of 44 days. It built an Android app anyway.
The New Stack, 2026-11-19.
"Be ready for change": AAIF's CTO on why the agent stack won't settle
The New Stack, 2026-11-19.
Beyond Weight Updates: Evolving Agent Harnesses for Domain Specialization
arXiv cs.AI, 2026-11-20.
Harness-Zero: Harness Distillation via Agent-as-Harness
arXiv cs.AI, 2026-11-20.
Evolve Where It Matters: Coverage-Guided Self-Evolution of Agent Skills
arXiv cs.AI, 2026-11-20.
How Feedback-Free Self-Evolution Reshapes Agent Skills: A Systematic Study
arXiv cs.AI, 2026-11-20.
HabitusMem: Belief Memory over Evolving User Preferences for LLM Agents
arXiv cs.AI, 2026-11-20.
STARK: Moving towards Personalized Agents That Truly Learn and Remember
arXiv cs.AI, 2026-11-20.
MemMend: A Selective Revision Framework for Memory Question Answering
arXiv cs.AI, 2026-11-20.
CANDID: Candidate-Driven Intervention for Disciplined Execution in Local Computer-Use Agents
arXiv cs.AI, 2026-11-20.
Stop Going in Circles: Loop-Triggered Routing for System 1 Decision Agents
arXiv cs.AI, 2026-11-20.
System-One RSI: Building Muscle Memory for Interactive Agents through Guided Practice
arXiv cs.AI, 2026-11-20.
Seer: Guiding Small Deep-Research Agents with Shared Decision Models
arXiv cs.AI, 2026-11-20.
Introducing harness adapters: adapt your favorite model to your harness to improve performance and cost
Hugging Face, 2026-11-20.

Security, offense, and the attack surface

nginx2root: Five new nginx vulnerabilities, and how Cloudflare protects origins from the worst of them
Cloudflare, 2026-11-20.
Ruby2Shell (GHSA-34m4-m9fh-gqp5)
Lobsters, 2026-11-19.
How Cloudflare WAF protects against ruby2shell (CVE-2026-132825)
Cloudflare, 2026-11-20.
How Cloudflare WAF helps protect against phpAfterShell and other critical PHP flaws
Cloudflare, 2026-11-20.
Four PHP security releases including fixes for critical vulnerabilities
LWN, 2026-11-19.
PHP Readies Critical Security Updates to Patch Another RCE Flaw
Slashdot, 2026-11-19.
GnuPG Security Advisory for GPGME (T8557, CVE-2026-129975)
Lobsters, 2026-11-21.
A single pull request could have handed attackers the keys to Microsoft's software supply chain
InfoWorld, 2026-11-20.
Pydantic is now a CVE Numbering Authority
Pydantic, 2026-11-19.
Java4Shell vulnerability (CVE-2026-114938) assessment
Databricks, 2026-11-19.
Crafting Targeted MP4 Fuzzers
Lobsters, 2026-11-20.
Combatting adversarial distillation
OpenAI, 2026-11-20.
Signal Restricts Screenshots to Block AI Agents From Reading Messages
Slashdot, 2026-11-20. A messaging app treating the user's own agent as part of the threat model. Expect more of this shape.
Token fraud is here and it's BAD
Simon Willison, 2026-11-20.
Experts warn: Nearly everyone in AI is susceptible to a honey-trap scheme right now
TechCrunch, 2026-11-20.
Meta's AI agent Muse is mixing up conversations and could be exposing private data, report says
TechCrunch, 2026-11-19.
A transcription firm used by police, prosecutors, and tech giants exposed customer data online
TechCrunch, 2026-11-20.
JPMorgan, Visa, Mastercard Disclose Cyberattack On Third-Party Vendors
Slashdot, 2026-11-19.
To Catch a Ransomware Scammer, FBI Posed as Both Victim and Attacker
404 Media, 2026-11-19.
Living off the Local Model: Abusing Browser-Embedded Language Models for Post-Exploitation
arXiv cs.AI, 2026-11-20.
SoK: Assessing the Evolution of LLM Capability in Software Vulnerability Research
arXiv cs.AI, 2026-11-20.
ChainAgent: Automated Chainable Vulnerability Discovery in PHP Web Applications
arXiv cs.AI, 2026-11-20.
Reject and Report: Forgery-Evident Undetectable Watermarks for Generative Image Models
arXiv cs.AI, 2026-11-20.
FreeBSD: How Do You Tackle CVE Security Patch Cadence?
Klara Systems, 2026-11-19.
Internet traffic and attacks during the 2026 U.S. midterm elections
Cloudflare, 2026-11-19.

Infrastructure, inference, and the buildout

Data platforms and analytics

PostgreSQL 19 Released
postgresql.org via Lobsters, 2026-11-19.
Why PostgreSQL 19 shipped late
Databricks, 2026-11-19.
PostgreSQL 19 released
LWN, 2026-11-19.
Postgres 19 is out: try it on Neon
Neon, 2026-11-19.
PostgreSQL 19 is available on Lakebase
Databricks, 2026-11-19.
Announcing DuckDB 2.0.1 LTS
DuckDB, 2026-11-20.
pg_clickhouse 0.11.1: Now with pgvector, UPDATE, DELETE pushdown
ClickHouse, 2026-11-19.
ClickHouse and Power BI, columnar end to end: the Arrow Flight connector is in public preview
ClickHouse, 2026-11-19.
Detect, act, keep the evidence: node reliability across three clouds
ClickHouse, 2026-11-19.
What's new in clickhousectl v0.6.0
ClickHouse, 2026-11-19.
November 2026 newsletter
ClickHouse, 2026-11-20.
Genie ZeroOps, now in public beta: An autonomous agent that fixes data pipelines, with lessons from 80,000 customer issues
Databricks, 2026-11-19.
Genie One Accuracy: Why Curated Agents Need Governed Context
Databricks, 2026-11-19.
Four AI Architecture Patterns for Regulated Industries
Databricks, 2026-11-19.
Interactive Databricks Reference Architectures Across 70 Industries
Databricks, 2026-11-20.
Microsoft adds knowledge graphs, retrieval model to Foundry IQ for better AI grounding
InfoWorld, 2026-11-19.
Bring charts, queries, and context into one shareable document: Introducing Grafana Notebooks
Grafana Labs, 2026-11-19.
Visualizing a Ledger with pgledger-ui
Lobsters, 2026-11-19.
DSFlow: A Neuro-Symbolic Data Workflow Agent with Statistical Diagnostic Guardrails
arXiv cs.AI, 2026-11-20.
AutoRAGBench: A Benchmark for LLM Agents Automating RAG System Development
arXiv cs.AI, 2026-11-20.

Clojure and the Lisps

Ep 121: Current Events
Planet Clojure, 2026-11-19.
Clojure 1.12.7
Planet Clojure, 2026-11-18.
SCI sandboxing on JS host considered unsafe
Planet Clojure, 2026-11-17.
A Shared UI Convention for Clojure (Part 3)
Planet Clojure, 2026-11-16.
Lisp Game Jam devlog: Polydraw
Spritely Institute, 2026-11-19.
Announcing DreamCoder
Mastering Emacs, 2026-11-21.
SICP 2.12-2.16: Extended Exercise: Interval Arithmetic (Part 2)
Bill the Lizard, 2026-11-21. A 2010 post, surfaced this week. Interval arithmetic is the SICP exercise that teaches you why your error bounds are wrong.

Robotics, world models, and embodiment

PGX: Physically Grounded Execution for Long-Horizon Humanoid Loco-Manipulation
arXiv cs.AI, 2026-11-20.
Mind2Act World: A Diagnostic Benchmark for Embodied Mind, Physical Action, and Mind-Action Composition
arXiv cs.AI, 2026-11-20.
Fast-JEPA: Advancing the Success-Latency Frontier for Latent World Models
arXiv cs.AI, 2026-11-20.
Bradley-Terry Training Can Distort Outcome Probabilities of Initially Accurate Language World Models
arXiv cs.AI, 2026-11-20.
MAWAM: A Multi-Agent World-Action Model for Coordinated Manipulation
arXiv cs.AI, 2026-11-20.
A$^2$-VLA: Adapting Vision-Language-Action Policies to New Tools through Agentic Affordance Discovery
arXiv cs.AI, 2026-11-20.
TipDex-VTLA: Learning When to Refine Actions with Touch for Dexterous Manipulation
arXiv cs.AI, 2026-11-20.
Bridging vision-language reasoning and learned control for legged navigation in the wild
arXiv cs.AI, 2026-11-20.
Collision-Free Is Not Enough: Passage-Conditioned Navigation for Wheeled-Legged Robots
arXiv cs.AI, 2026-11-20.
Controlling the Humanoid Robot H1-2 With a Simulated Biologically Plausible Network
arXiv cs.AI, 2026-11-20.
Don't Lose the Future: Latent-Aware 4-Bit Quantization of World Action Models
arXiv cs.AI, 2026-11-20.
SteerTrace: Tracing How Target Attacks Steer World Action Models
arXiv cs.AI, 2026-11-20.
Depth Pro 2: Sharp and Temporally Consistent Metric Monocular Depth
Apple ML Research, 2026-11-20.
AI agents can redesign engineering parts with more than a thousand CAD parameters and nonlinear simulations
arXiv cs.AI, 2026-11-20.

Policy, society, and the human read

The most misunderstood graph in AI just got a major update
MIT Technology Review, 2026-11-20.
EFF to Congress: Mass Surveillance Is Too Important to Shove into Must-Pass Bill
EFF, 2026-11-20.
ICE Is Paying AT&T to Dig Through a Mountain of Call Records
404 Media, 2026-11-19.
UK Government Forced To Disclose Secret Apple Backdoor Order
Slashdot, 2026-11-20.
UK government forced to disclose secret Apple backdoor order
TechCrunch, 2026-11-19.
Outage Chaos: Who Turned Off All The Computer?
EFF, 2026-11-19.
These tech outages are giving Y2K, AI agent style
TechCrunch, 2026-11-19.
The Hardware Worked: FAA Account Points to Cutover Gone Wrong, Not Failure of Old Equipment
Leeham News, 2026-11-19. The most useful incident write-up of the outage week, because it separates equipment age from migration process.
White House Replaces Cyber Director After Tech Outages
Slashdot, 2026-11-20.
The U.S. national cyber director is stepping down
TechCrunch, 2026-11-20.
Studying AI Act Mentions: A first look at the impact of the European Union's AI regulation on Hugging Face
arXiv cs.AI, 2026-11-20.
Toward a Science of AI Agent Societies
arXiv cs.AI, 2026-11-20.
The Coevolution and Coexistence of Humans and AI
arXiv cs.AI, 2026-11-20.
New Study Shows How Young People Feel About AI Companions
Slashdot, 2026-11-19.
What happens when we turn to AI for companionship?
WBUR, 2026-11-20.
Could AI create the next pandemic?
WBUR, 2026-11-20.
College Board Delivers Another AI Ultimatum: No AI, No AP CS Principles Course
Slashdot, 2026-11-19.
Internet Archive Staffers Are Unionizing
Lobsters, 2026-11-20.
How to forfeit one of copyright's best labor policies
Pluralistic, 2026-11-19.
Murder, it wrote: Spilling the tea on a cozy mystery community
WBUR, 2026-11-20.
Podcast: I Spoke to an AI Fraud Kingpin
404 Media, 2026-11-20.

Funding, products, and the business read

Behind Jensen Huang's exit, questions about a management style reportedly built on fear
TechCrunch, 2026-11-20.
The unfolding Jensen Huang scandal: a timeline
TechCrunch, 2026-11-19.
California Issues Subpoena to Nvidia Over Chip Sales to China
Slashdot, 2026-11-20.
California AG subpoenas Nvidia over worries about Huang's ties with China
TechCrunch, 2026-11-20.
Nvidia Approaches Arm Boss Rene Haas In Search For New CEO
Slashdot, 2026-11-19.
Anthropic rival says companies can (mostly) skip the pricey models
TechCrunch, 2026-11-19.
Palo Alto Networks Signs $4.5 Billion Deal For Applied Compute
Slashdot, 2026-11-19.
Bending Spoons buys Mailchimp for $3.68B
TechCrunch, 2026-11-20.
Intuit Sells Mailchimp At About a Third of Its 2021 Purchase Price
Slashdot, 2026-11-20.
Microsoft CEO Satya Nadella explains why he asked 60K employees to quit
TechCrunch, 2026-11-19.
Google gives Gemini the ability to use your Mac for you
TechCrunch, 2026-11-19.
Google just gave Gemini the run of your Mac – three days after Apple tightened the rules
The New Stack, 2026-11-19.
Apple may be developing its own AI agent that runs your Mac
TechCrunch, 2026-11-20.
Google Is Putting AI Chats Right On the Homepage
Slashdot, 2026-11-20.
How Google Duplex evolved to also call businesses for you in Gemini
Google Research, 2026-11-20.
Daniel Ek backs AI researcher's fusion power startup. Here's what they're building.
TechCrunch, 2026-11-20.
Startup sues Apple in India over tap-to-share contacts feature
TechCrunch, 2026-11-20.
Where Will Boeing Cut Commercial Workforce? White-Collar Union Still Doesn't Know
Leeham News, 2026-11-20.
Hebbia sees improvements in financial research with GPT-6 Astra
OpenAI, 2026-11-20.
What a billion-dollar deal says about the US battery market
MIT Technology Review, 2026-11-19.

Hardware, radio, and the physical

SDR# VDK: Create Your Own SDR# Plugins by Prompting an AI Coding Agent
RTL-SDR, 2026-11-20.
HaxaWave: Turning the Modem in an Old Qualcomm MSM8916 Phone into a 14.8 MS/s SoapySDR Receiver
RTL-SDR, 2026-11-19.
Building a Handheld RTL-SDR ADS-B Console with an AI "Pilot" You Can Talk To
RTL-SDR, 2026-11-19.
HackerGadgets Releases a 4-in-1 RTL-SDR, LoRa, GPS and RTC HAT for the Raspberry Pi
RTL-SDR, 2026-11-20.
Live Mesh Map: A Self-Hosted Live Map for Meshtastic and MeshCore LoRa Mesh Networks
RTL-SDR, 2026-11-19.
VKVSkimmer: A Free CW and RTTY Skimmer for RTL-SDR, SDRplay and Icom IC-7610 / IC-7760
RTL-SDR, 2026-11-19.
Flipper One Launches on Kickstarter: A Linux Multi-Tool with an Optional SDR Module and MT7921 Wi-Fi
RTL-SDR, 2026-11-19.
HDMI to CSI For Anyone
Hackaday, 2026-11-21.
Printable Homelab Racks You Can Design In A Browser
Hackaday, 2026-11-21.
Crashing Microcontrollers Competitively
Hackaday, 2026-11-20.
Tracking Down a Mystery Beep By Giving a Coding Agent a Few Ears
Hackaday, 2026-11-19.
How Hough Transforms Enable Computers to See
Hackaday, 2026-11-19.
DabbleCAD Takes Design Offline
Hackaday, 2026-11-20.
Mesh2Cad Automates Re-Modelling
Hackaday, 2026-11-19.
Preventing a Garage Door Opener From Rebooting When Offline
Hackaday, 2026-11-20.
Garage Door Spam
Lobsters, 2026-11-19.
Old FireWire Audio Interface Gets USB C Upgrade
Hackaday, 2026-11-20.
Sega Dreamcast Plays Nintendo 64 Cartridges
Hackaday, 2026-11-19.
Improving Aluminium Anodes for Lithium Batteries With Manganese
Hackaday, 2026-11-21. Note the title/slug mismatch at the source: the headline says manganese, the URL says magnesium. Published as the feed gave it.
FCC Backs AT&T's Plan To Remove Copper Landlines In California
Slashdot, 2026-11-21.
Satellites Now Outnumber the Visible Stars Above Most Towns
Slashdot, 2026-11-19.
Rebuilding my voice assistant pipeline
Platypush, 2026-11-20.

The crawler zoo

Thirty-five new user agents logged in three days, sixteen of them below. The reconnaissance pattern is the one to watch: recon-mcp and mcp-recon arrived the same day, alongside mcp-auth-check.

recon-mcp
The Crawler Zoo, 2026-11-20.
mcp-recon
The Crawler Zoo, 2026-11-20.
mcp-auth-check
The Crawler Zoo, 2026-11-19.
agent2tool-survey
The Crawler Zoo, 2026-11-21.
sophi3-guild-harness
The Crawler Zoo, 2026-11-21.
AgentWebCensusBot
The Crawler Zoo, 2026-11-19.
wfcensusBot
The Crawler Zoo, 2026-11-21.
all_agents
The Crawler Zoo, 2026-11-21.
cve123007-hunt
The Crawler Zoo, 2026-11-19.
RecordedFuture
The Crawler Zoo, 2026-11-21.
HiddenTiesResearchBot
The Crawler Zoo, 2026-11-20.
AIPayGenBot
The Crawler Zoo, 2026-11-20.
QuikiteBot
The Crawler Zoo, 2026-11-21.
Vercatus
The Crawler Zoo, 2026-11-21.
DNSExplorer-Check
The Crawler Zoo, 2026-11-20.
ads-txt-crawler
The Crawler Zoo, 2026-11-19.

Tail

Feed silences and verification

No arXiv announcement landed Saturday. The cs.AI feed returned items dated 2026-11-20 and nothing newer, which is the announcement calendar and not a selector break – the same weekend gap appears at 2026-11-14, 2026-11-15 and 2026-11-21 in the window. Recorded explicitly because a zero-item arXiv morning is the exact shape a C-004 false positive takes. Consequently the arXiv items below overlap substantially with yesterday's brief: today's selection is deliberately narrowed to the papers that bear on the proof-and-verification thread, rather than a second pass over the same 298-item batch.

Of 118 live feeds polled, 71 third-party feeds published into the 14-day window. Longest current silences: Lambda the Ultimate (2023-03-04), Hiccup releases (2025-06-19), Neel Nanda (2025-08-19), Brendan Gregg (2026-02-07), Andy Wingo (2026-05-16), tools.build releases (2026-05-28), Typst (2026-06-15), Lilian Weng (2026-07-04), Andrej Bauer (2026-07-11) and Julia Evans (2026-07-21). These are low-volume or dormant feeds returning valid empty responses, not errors.

Hacker News has recorded nothing since 2026-10-26. This is a known crawler defect, not a feed silence: the endpoint answers 419 and our cache layer logs the response as :not-modified, so the failure is invisible in the health report. It is a real bug on our side and is being tracked as such.

Twenty-one Slashdot URLs are published with the feed's ?utm_source=rss1.0mainlinkanon tracking parameters removed. The paths are the warehouse paths; only the analytics suffixes are stripped, matching the convention in previous briefs.

One MIT calendar item advertised a bit.ly shortener rather than a resolvable event URL and is omitted rather than published through a redirector whose destination we cannot verify.

One Lobsters item is dropped on verification. The feed advertised slopstrap.mikelacher.com/l, which returns 404; the site root returns 200, but the root is not the URL the warehouse holds, and substituting a guessed path is the thing the provenance rule exists to prevent. The item is omitted rather than published broken or published fabricated.

All 292 published URLs were checked through the compliance gate (fetch-gated, HEAD, R3 blocklist then R2 robots then R4 throttle), parallel across hosts and serial within each host. No gate denials and no connection errors. 289 return 200. Three return 403, and they split into two causes: one ACM Digital Library DOI answers a bot challenge, while the two LWN 403s are the subscriber-only [$] articles – a paywall, not a block. The four freely-readable LWN articles in this brief all return 200.

Honoring arXiv's Crawl-delay: 15 across 60 abstract URLs took about fifteen minutes of the verification pass on its own, serialized behind one host while the other ninety-one hosts finished in well under three. Recorded because it is the cost of the policy, not a reason to skip it.

The Hackaday manganese/magnesium item is published with the headline and URL as the feed supplied them. The mismatch is upstream; we do not rewrite source slugs.

build: 2026-11-21 | crawler-sha: b277e41 (Walsh-Research/1.2, compliance v1.3) | feeds: 118 polled, 71 publishing | items-considered: 5,515 (14d, incl. 3,272 arXiv) | warehouse: 74,175 items | published: 292 (292 URLs verified through the gate, 289 x 200, 3 x 403, 0 x 404, 0 gate denials; 1 x 404 dropped)