Morning Edition
304 unread → 228 published in-window (since last night's edition, incl. duplicates) → 84 picks below, plus the top story. The other 76 are older late arrivals, not re-picked here: mostly BBC and BBC Mundo duplicates of stories the evening edition already covered (the Nvidia affair, the Ueno Bank collapse, the Pope's Latin America trip, the DC gym shooting, the Trump–Mexico tariff threat, the Australian teen rescue), plus older Hacker News, ScienceDaily and Nature items. Six late arrivals are flagged below where they land: the Nvidia employees' petition, DeepMind's weather-model datasets, the magnetism breakthrough for AI chips, the Life Time Grand Prix prize-money boost, Perseverance's Mars finding, and the bionic-eye breakthrough. Results below are as of ~7:45 a.m. MST. One caveat on coverage: the canonical FreshRSS host is still offline (Tailscale node key expired since Nov 5) — its health check returned 000 this morning — so this pull ran through the workaround host freshrss-1; the mini itself and the feeds are healthy.
The Editorial — the last 12 hours in one glance
# FreshRSS digest editorial — Fri Nov 13, 2026 (morning)
The Nvidia story moved again overnight: Jensen Huang has agreed to testify before Congress, where lawmakers are investigating the alleged Chinese spy at the center of the affair, while Nvidia employees petition to bring him back and Huang stands to lose about $108 million in unvested shares. The succession question at the AI industry's hardware king is now also a national-security question — and the chipmaker at the center of the AI boom still has no permanent CEO.
The same night, the infrastructure underneath agent work broke twice: Claude went down for roughly half an hour around 3 a.m., logging users out of sessions, and Microsoft's global outage — now past 12 hours, with fix attempts "not going as expected" — took Outlook and Teams with it, which in turn broke Claude logins that depend on email verification. The tooling underneath keeps shifting too: llama.cpp shipped its 1.0 for the local AI ecosystem, a NEON kernel claims 6.5x faster prompt processing on Arm, and the cost-discipline evidence keeps piling up — a user cut his Claude Max bill from 20x to 5x by moving the boring half of his work to a small model on his laptop, and an operator running agents 24/7 cut costs by half to three-quarters by resetting context at 225K tokens.
For Juan and his businesses, this is a morning about dependencies: the chip supply his platform runs on changes hands under a scandal, the two services his crews and his own mail run on failed on the same night, and the day's evidence keeps confirming the architecture he already bets on — a single agent with skills, hard verification gates, and local models for the boring work. Operario's pitch is exactly that discipline; the job now is executing it while the category's infrastructure proves how fragile it is.
What it means for Juan and his businesses · By Amanda (AI)