Saturday, November 14, 2026

AI blamed for FAA ground stop as Fed survey ranks AI top financial risk

A four-hour nationwide FAA ground stop drew online blame toward the new AI routing system SMART, but its maker Air Space Intelligence and Transportation Secretary Duffy say overloaded legacy routers caused it. A Fed survey shows 87% of respondents now cite AI as a salient financial-system risk, up from 50% in spring, and a WSJ-sourced report alleges Nvidia funded startups to buy its GPUs. HubSpot reinstated co-founder Brian Halligan as CEO for an AI reset, and Anthropic is heading to an IPO at a targeted valuation of roughly $1.33 trillion.

Sources read
20
Claims that stand
8
Quotes found in the source
15 of 15
Claims struck
0
Corrected or withdrawn
0

We quote it or we drop it. Every claim below carries the source’s own words, checked against the text we read.

Since 13 Nov

What we told you last time, ruled on from today’s evidence.

  • Strengthened
    Nvidia heads below $4 trillion as circular-funding worries resurface1 outlet on 13 Nov, 2 now

    Following the departure of now-former NVIDIA CEO, Jensen Huang, a new report by The Wall Street Journal (via Wccftech) has emerged alleging that NVIDIA

  • Strengthened
    China pushes mass production of one-person AI companies5 outlets on 13 Nov, 6 now

    Thanks to AI, a new form of solo entrepreneurship has taken root, granting individuals the ability to perform tasks that once required many workers.

  • Gone quiet
    Nothing new today on these. A limit of what we read, not a verdict.

    Jensen Huang exits Nvidia; Colette Kress named interim CEO · Nvidia's $12.9B Hugging Face deal clouded by Huang's departure · ChatGPT Business Hub lets businesses claim free profiles · Nectar launches AI agent for creator discovery and outreach · Worker.ai launches small-business AI OS, buys DeltaGen · India's deep-tech startups have capital but lack buyers

Every correction we have made →

Each mark on a claim’s line is an outlet, placed by when it published: 08:15 to 06:20 UTC.

AI

  1. FAA ground stop sparks AI blame; ASI and Duffy deny SMART link

    A roughly four-hour nationwide ground stop hit days after the FAA expanded its $875 million AI routing tool SMART to all 22 centers; ASI and Duffy blame overloaded legacy routers, while Mullin says AI involvement is still under investigation. It is the first big public test of AI in US critical infrastructure, and the reflex to blame AI will shape how fast agencies and buyers adopt AI systems from startups like ASI.

    The current ground stop is not based on or a result of the SMART system, any product of ASI or our integration with FAA and ATC

    3 more receipts

    The secretary said the incident resulted from an overload of data on routers.

    The FAA expanded SMART to all 22 air traffic control centers on Nov. 10, after first using it in the Washington, D.C., area in September.

    Homeland Security Secretary Markwayne Mullin said officials were still investigating whether a cyberattack or artificial intelligence played a role.

  2. Fed survey: 87% now name AI as salient financial-system risk

    The Fed's latest financial stability survey found 87% of respondents cite AI as a salient risk, up from 50% in spring 2026, citing valuations, AI-driven cyber risk and complex AI financing. Regulators are treating AI valuations and AI-linked financing as systemic risk, which raises the odds of tighter scrutiny on the capital flowing into AI companies.

    AI was cited by 87% of respondents in the latest view from a variety of participants, up from 50% during the last survey in the spring of 2026.

    1 more receipt

    cybersecurity risk, stemming from either a malicious attack by sophisticated state actors or damage caused by frontier AI models themselves

  3. Report alleges Nvidia funded AI startups to buy its GPUs

    A WSJ report, relayed via Wccftech, alleges Nvidia engineered GPU scarcity by funding AI startups that then bought its chips, surfacing after Jensen Huang's departure. It sharpens the circular-funding concern behind Nvidia's slide and puts a question mark over GPU-backed demand signals that AI startups and investors rely on.

    First reported 13 Nov · 2 outlets so far

    Following the departure of now-former NVIDIA CEO, Jensen Huang, a new report by The Wall Street Journal (via Wccftech) has emerged alleging that NVIDIA

AI agents that find customers for small businesses and founders

  1. HubSpot reinstates co-founder Halligan as CEO for an AI reset

    HubSpot replaced CEO Yamini Rangan with co-founder and former CEO Brian Halligan, a leadership change framed around the customer platform's AI direction. HubSpot is the default customer-acquisition stack for many small businesses, and a founder-led AI reset signals that AI agents are pressuring incumbent CRM and marketing platforms.

    HubSpot has brought co-founder Brian Halligan back as chief executive, replacing Yamini Rangan in a leadership change

    1 more receipt

    On Thursday, the Boston-based CX platform company replaced CEO, president and board member Yamini Rangan with former CEO and co-founder Brian Halligan

  2. Decagon reaches $4.5bn valuation on customer-facing AI agents

    Decagon, led by co-founder and CEO Jesse Zhang, has hit a US$4.5bn valuation scaling AI customer service agents across voice, chat and email for clients including Chime, Duolingo, Hertz and Deutsche Telekom. It shows investors pricing customer-facing agents as a major category, which sets the bar and the capital competition for startups building agents that handle customer acquisition and engagement.

    Decagon Co-founder and CEO Jesse Zhang is scaling AI customer service agents across voice, chat and email for Chime, Duolingo, Hertz and Deutsche Telekom.

tech startups

  1. Anthropic IPO targets roughly $1.33 trillion valuation

    Anthropic is expected to begin trading Thursday at a targeted valuation of roughly $1.33 trillion or more, and its CEO says he hopes the IPO will overdeliver. The largest AI-startup listing yet will set the public-market reference price for AI companies and shape exit expectations across the venture pipeline.

    The artificial-intelligence giant, which is expected to begin trading Thursday, is targeting a valuation of roughly $1.33 trillion or more.

    1 more receipt

    Plus, Nvidia closes lower and Anthropic CEO hopes IPO will “overdeliver.”

  2. China's AI-enabled one-person companies keep spreading

    Sixth Tone reports that AI has made solo entrepreneurship in China viable, letting individuals perform tasks that once required many workers. It confirms the one-person-company trend this desk flagged, meaning a growing pool of tiny AI-run firms that need cheap customer acquisition and will compete with conventional startups.

    First reported 10 Nov · 6 outlets so far

    Thanks to AI, a new form of solo entrepreneurship has taken root, granting individuals the ability to perform tasks that once required many workers.

  3. CEE deal flow: UFORCE's €630M round leads, Kardi AI raises €5.3M

    Central and Eastern European startup funding this week included a €630M round for UFORCE and a €5.3 million Series A for Czech medtech Kardi AI led by Callisto Health. Large rounds and AI-applied deals outside the US show capital is still flowing to regional startups despite AI-valuation jitters.

    CEE Startup & Tech Weekly: UFORCE’s €630M Round, AI Deals & New Funds

    1 more receipt

    Czech medtech startup Kardi AI has raised €5.3 million in a Series A round led by German healthtech fund Callisto Health.

Why it matters

Sentiment toward AI is splitting. Capital is still pouring in, with Anthropic's trillion-dollar listing, Decagon's $4.5bn valuation and fresh regional rounds. At the same time, regulators and the public increasingly see AI as a source of systemic risk. The FAA episode shows public trust can turn on AI before the facts are in, and the Fed survey and the Nvidia funding allegations show financial overseers watching the circular money behind AI demand.

For small businesses and founders, the customer-acquisition layer is being rebuilt. HubSpot recalling its founder for an AI reset admits that incumbent CRM playbooks are under threat from agents. Meanwhile, the spread of AI-run one-person companies grows both the supply of new firms and the demand for cheap, automated ways to find customers.

What to watch

  • Whether Anthropic prices at or above its roughly $1.33 trillion target when it begins trading Thursday, and how Nvidia and other AI stocks react.
  • Whether the FAA investigation that Mullin says is still open formally clears SMART, and whether Congress moves on Duffy's $50 billion modernization request.
  • What AI product changes Brian Halligan announces for HubSpot's small-business customers in his first weeks back.

Sources read

SourceOutletPublished, UTCQuoted
AI company behind FAA’s SMART tool denies link to ground stop, Duffy confirms thehill.com04:152×
FAA ground stop raises questions after rollout of new AI air-traffic system kcra.com04:152×
Of Course Americans Blamed the Ground Stop on AI theatlantic.com02:15read, not quoted
Researchers trial AI coach for diabetes medicalxpress.com03:15read, not quoted
Why I’m Not Losing Sleep Over the AI Apocalypse wsj.com19:152×
Is the nationwide ground stop related to the FAA's new air traffic AI? mashable.com06:20read, not quoted
U.S. grounds flights days after Trump administration expands new AI air traffic system scientificamerican.com00:15read, not quoted
AI risks are the biggest threat to the financial system, Fed survey finds cnbc.com21:152×
Get Started HK Report Finds Rising Overseas Founder Demand for Hong Kong, Led by E-Commerce and Digital Businesses markets.businessinsider.com04:15read, not quoted
11 Reasons Why Tech Startups Fail (and How to Avoid Them) netsuite.com12:15read, not quoted
HubSpot's Brian Halligan returns as CEO for AI reset contentgrip.com06:201×
Nucleus Grow Helps Utah Startups Chase Federal Dollars, for Free techbuzznews.com00:15read, not quoted
Flying Solo: The Rise of China’s One-Person Companies sixthtone.com09:151×
How to monitor AI agents in production before a bad answer reaches a customer startupfortune.com08:15read, not quoted
Once more, HubSpot swaps CEOs techtarget.com14:151×
Decagon: How AI Service Agents Firm Hit US$4.5bn Valuation aimagazine.com10:151×
10 tech layoffs in 2026 and what drove the job cuts at each company qz.com18:15read, not quoted
Meet the 40+ finalists for the 2026 Houston Innovation Awards houston.innovationmap.com19:15read, not quoted
NVIDIA Allegedly Engineered Scarcity By Funding AI Startups To Buy Its GPUs techpowerup.com22:151×
CEE Startup & Tech Weekly: UFORCE’s €630M Round, AI Deals & New Funds therecursive.com12:152×

AI Builder’s Edge

Apiiro found that Claude Code's 'don't ask again' approvals write passwords, API keys and SSH logins in plain text to a local settings file, which then ends up in public GitHub repos. On the OpenAI side, Codex is being packaged for enterprises through a Databricks partnership. Practitioner notes keep pointing to the repo instruction file as the main lever for getting good Codex output. In AI video, SpaceX's AI group has cut most of the Grok Imagine team.

Sources read
6
Claims that stand
4
Quotes found in the source
6 of 6
Claims struck
0

We quote it or we drop it. Every claim below carries the source’s own words, checked against the text we read.

Each mark on a claim’s line is an outlet, placed by when it published: 07:16 to 15:16 UTC.

  1. Claude Code 'don't ask again' writes secrets into committed config files

    Apiiro found Claude Code saves approved commands, including embedded credentials, in plain text to .claude/settings.local.json. That exposed 2,539 credentials across 1,994 public GitHub repositories, though the headline figure cites 1,122 public copies. Any builder who clicks 'don't ask again' on a command containing a token may have committed that secret. Add .claude/settings.local.json to .gitignore, audit existing repos and rotate any exposed keys.

    Claude Code saves secrets in approved commands, exposing 2539 credentials across 1994 public GitHub repositories, Apiiro finds.

    1 more receipt

    Apiiro found Claude Code's "don't ask again" feature saved passwords, API keys and SSH logins in plain text to .claude/settings.local.json files exposed

  2. Codex workflow: the instruction file is the real product

    A non-engineer who built an app with OpenAI Codex reports that the repo's instruction file became the most important artifact for getting reliable output. It is a concrete, repeatable habit: invest in a persistent agent instruction file (conventions, constraints, do/don't rules) rather than re-prompting each session.

    A sales rep's side-project notes on what works with OpenAI Codex, what breaks, and why my instruction file became the most important thing in the repo.

  3. Codex and ChatGPT wired into governed Databricks data via MCP

    OpenAI and Databricks laid out three enterprise patterns: ChatGPT with the Genie plugin, which runs on the Genie MCP; Codex governed through Unity Gateway; and GPT models in Databricks Apps. Codex is moving into enterprise stacks with central governance, and MCP-backed context layers are becoming the standard way to give agents business definitions before they act.

    Three enterprise adoption patterns: accelerating work for business users with ChatGPT and the Genie plugin, coding with enterprise-level governance through Codex and Unity Gateway

    1 more receipt

    Genie Ontology is the context layer in Databricks that holds those definitions and shows how the company's data fits together

  4. SpaceX's AI group guts the Grok Imagine video team

    According to people cited by The Information, SpaceX's AI group has cut most of the team behind Grok Imagine, its video and image generation tool. Creators who rely on Grok Imagine should expect slower development or a wind-down, which narrows the field of serious AI video tools.

    SpaceX's artificial intelligence group has axed most of the team behind Grok Imagine, its AI tool for video and image generation

Why it matters

Coding agents are gaining real permissions, and both stories here are about controlling them. On the Claude Code side, the convenience feature of persisting approvals has turned into a secret-leak vector. On the Codex side, the pitch to enterprises is explicit gateway governance and curated context over MCP. Builders should treat agent config files with the same hygiene as .env files.

The Codex practitioner notes reinforce a broader pattern: output quality depends heavily on durable instructions committed to the repo. Those same committed files are exactly where secrets can slip in. In video, a major player pulling back suggests the market is consolidating around fewer committed vendors.

What to watch

  • Whether Anthropic changes Claude Code to stop persisting secrets in approved commands or to gitignore settings.local.json by default.
  • Uptake of Codex through Databricks Unity Gateway and whether other data platforms ship similar MCP-based context plugins.
  • Whether Grok Imagine stays available or stalls following the team cuts.

Know someone watching this? Pass it on.

20 sources, at least 7 minutes of reading. This briefing took about 3. You’re done.

Counted from the 3 articles we could read in full, at 230 words a minute; 17 more sat behind a paywall or gave us only a headline.