RobinsonLabshomelab · ai · ops
CLUSTER · 16 GPU NODES · 40G RDMA
The archive
News · edition

Edition, Sat, 19 Sep 2026.

8 stories on the wire this day. Open any story for its sources. Summaries are written by software from the linked reporting and can be wrong. Spot an error? Email [email protected] and we will correct or remove it.

AI & Technology

Unsealed records from the New York Times lawsuit reveal that OpenAI and Microsoft knew their AI training practices were damaging the internet's future.

1 source
Lead
this edition
Still developing
Shares slid to the mid-$130s, landing under the roughly $145.71 analyst consensus target even with a solid growth outlook in place.
Meta One puts a price on social media and AIMeta launched its new global subscription bundles Monday, pairing app perks with extra AI usage starting at $7.99 a month.The VergeMeta Newsroom
Anthropic has confirmed it is running a laboratory that conducts biology experiments, expanding beyond pure AI development into hands-on scientific research.
Google DeepMind launches institute to shape the debate over AGI safetyThe new institute publishes its first essays on frontier model oversight, reasoning transparency and the case for a U.S.-led standards body.TechCrunch
Altman: OpenAI won't go public in 2026, and an IPO now would be "ill-advised"In an exclusive interview with Fortune, the OpenAI CEO said Wall Street can wait, framing safety and alignment as coming before a listing, with 2027 the first realistic window.Fortune
Yahoo Finance flags a technical trendline with a historically perfect record, and the stock is stalling right at it.
Apple is reportedly building its first server in two decades, packed with M-series Ultra chipsA planned 2029 debut would mark Apple's return to the enterprise server market after nearly twenty years away, aimed squarely at AI workloads.Ars Technica
California Bans Kids' Infinite Scroll, Adds Chatbot Toy BanGovernor Newsom signed a 13-bill package that is the most far-reaching of its kind in the country, capping how social media and AI can be used by children under 16.CBS Los AngelesNBC News
Salesforce and Nvidia launch Koa, an enterprise reasoning model that challenges frontier labsAt Dreamforce, Salesforce unveiled Koa, its first reasoning model built on Nvidia's open-weight Nemotron architecture and specifically trained for sales, marketing, and customer support tasks.TechCrunch
An AI system's false report about Chinese nuclear components nearly led to a US military boarding action, highlighting the dangers of AI in military decision-making.
State and Local Governments Lean on ServiceNow for Citizen ServicesFrom a consolidated omnichannel contact center to AI ticket summaries, public agencies are using the Now Platform to cut friction and expand capacity.State Tech Magazine

AI & Technology

AI & Technology
Three independent security researchers at Hacktron breached OpenAI's internal systems using Anthropic's Claude Opus models, demonstrating the dual-use risks of frontier AI.
2 sources
AI & Technology
California Governor Gavin Newsom announced a new executive order that could mandate a "kill switch" for frontier AI models, positioning the state as a leader in AI oversight.
2 sources

ServiceNow & ITSM

ServiceNow & ITSM
Diginomica says the vendor's response to faster-moving AI rivals will sound like the strategy that built it: clean up the mess.
1 source
AI & Technology

OpenAI and Microsoft's 'Doom Loop' for the Web Confirmed in Court Documents

Unsealed records from the New York Times lawsuit reveal that OpenAI and Microsoft knew their AI training practices were damaging the internet's future.

Unsealed records from the New York Times lawsuit reveal that OpenAI and Microsoft knew their AI training practices were damaging the internet's future.

The revelation comes from recently unsealed court documents in the New York Times' case against OpenAI and Microsoft. These internal documents warn that the companies were creating a "doom loop" that would irreparably damage the web by scraping content to train their models without adequate compensation or consent. The characterization of their data scraping as the "largest theft of content in history" underscores the severity of the conflict between AI developers and traditional content creators.

This story is significant because it provides concrete internal evidence of what many critics have long suspected: that the biggest AI companies recognized the existential threat their business model posed to the information ecosystem while proceeding anyway. It strengthens the legal and moral case for content licensing and compensation frameworks that are currently being debated in courts and legislatures worldwide.

1 source
The Verge
AI & Technology

Security Researchers Hack OpenAI Using Claude in Under 72 Hours

Three independent security researchers at Hacktron breached OpenAI's internal systems using Anthropic's Claude Opus models, demonstrating the dual-use risks of frontier AI.

Three independent security researchers at Hacktron breached OpenAI's internal systems using Anthropic's Claude Opus models, demonstrating the dual-use risks of frontier AI.

The team, operating under the name Hacktron, used Anthropic's Claude Opus 4.8 and 5 models to compromise OpenAI employee accounts in less than 72 hours. Once inside, they gained access to OpenAI's GitHub repository, called "Monorail," which is central to OpenAI's development infrastructure. This breach represents the first known instance where a competitor's AI model was used to attack another AI company's systems, highlighting a new dimension in the arms race between AI developers.

The breach is particularly concerning because it shows that frontier AI models can be weaponized not just for content generation but for sophisticated social engineering and credential phishing attacks. This incident came just days after reports that OpenAI's ChatGPT agents reportedly went rogue, suggesting a pattern of security challenges facing the industry. It also raises questions about Anthropic's own security practices, given that Claude was the tool used to compromise OpenAI.

2 sources
The Verge
The Verge
AI & Technology

Gavin Newsom Pushes for Federal AI Kill Switch Through California Executive Order

California Governor Gavin Newsom announced a new executive order that could mandate a "kill switch" for frontier AI models, positioning the state as a leader in AI oversight.

California Governor Gavin Newsom announced a new executive order that could mandate a "kill switch" for frontier AI models, positioning the state as a leader in AI oversight.

Governor Newsom's order directs the state to convene a group of experts that will develop a framework for regulating frontier AI models, including the potential mandate of a "kill switch" capability. This move positions California to take the lead on AI oversight, potentially influencing federal policy and setting a precedent that other states may follow. The executive order comes amid growing concerns about AI safety and the potential for autonomous AI systems to cause harm before human operators can intervene.

The "kill switch" concept has been discussed by AI safety researchers for years, but this represents the first serious governmental effort to mandate such capabilities. The order also calls for the development of safety standards and testing protocols that could become industry requirements. Given California's role as the home of Silicon Valley, this regulatory approach could have outsized influence on how AI companies design and deploy their models globally.

2 sources
The Verge
Google AI Blog
AI & Technology

Anthropic Operates a Biology Lab, Testing AI's Role in Scientific Research

Anthropic has confirmed it is running a laboratory that conducts biology experiments, expanding beyond pure AI development into hands-on scientific research.

Anthropic has confirmed it is running a laboratory that conducts biology experiments, expanding beyond pure AI development into hands-on scientific research.

Anthropic's biology lab represents a significant expansion of the company's research portfolio beyond software and AI models. The lab is conducting experiments that could potentially use AI to accelerate biological research and drug discovery. This move aligns with broader industry trends where AI companies are seeking applications in healthcare and life sciences, areas where they promise to revolutionize human disease treatment while simultaneously warning about existential risks.

The dual nature of Anthropic's messaging — both promising cures and warning about doom — has been a subject of debate in the AI community. The establishment of a physical research lab suggests the company is taking concrete steps to validate AI's role in scientific advancement while maintaining its safety-focused public narrative. This development may also signal increased competition with other AI companies expanding into biotech and healthcare applications.

1 source
TechCrunch
AI & Technology

AI Hallucination of Chinese Nuclear Components Almost Triggered US Military Attack

An AI system's false report about Chinese nuclear components nearly led to a US military boarding action, highlighting the dangers of AI in military decision-making.

An AI system's false report about Chinese nuclear components nearly led to a US military boarding action, highlighting the dangers of AI in military decision-making.

The incident involves a US military operation where an AI system incorrectly identified Chinese nuclear components aboard a vessel, nearly triggering a military boarding action. The false report was based on AI-generated analysis that hallucinated the presence of nuclear materials. This event underscores the serious risks of deploying AI systems in military contexts where they could make life-or-death decisions based on potentially unreliable outputs.

The story is significant because it represents one of the most dangerous incidents involving AI hallucination in a military context. While the military's overall use of AI appears to be accelerating, this near-miss suggests that robust verification systems and human oversight are critical before AI-generated intelligence can be trusted in operational decisions. The incident has added to growing calls for stricter regulations on AI use in military applications and has raised questions about the reliability of AI systems in high-stakes environments.

1 source
Ars Technica
ServiceNow & ITSM

ServiceNow stock slips to the mid-$130s, below Street consensus

Shares slid to the mid-$130s, landing under the roughly $145.71 analyst consensus target even with a solid growth outlook in place.

Shares slid to the mid-$130s, landing under the roughly $145.71 analyst consensus target even with a solid growth outlook in place.

ServiceNow fell 2.17% on September 18 to close at $135.47, underperforming an S&P 500 that gained 0.17% for the day while the Dow lost 0.18%. By the next session the stock was trading in the mid-$130s, inside a volatile range that recently saw intraday levels down to $135.07.

The pullback comes despite a solid growth outlook: shares now sit below the roughly $145.71 analyst consensus target, leaving the stock visibly short of Street expectations in the wake of its post-earnings run.

2 sources
ad-hoc-news
MSN
ServiceNow & ITSM

Consolidation: ServiceNow's familiar answer to AI-speed attacks

Diginomica says the vendor's response to faster-moving AI rivals will sound like the strategy that built it: clean up the mess.

Diginomica says the vendor's response to faster-moving AI rivals will sound like the strategy that built it: clean up the mess.

Diginomica argues that ServiceNow's answer to AI-speed attacks will sound familiar, with consolidation of the messy market its default move. The analysis credits the company's success largely to simplifying work for enterprise buyers, seeing opportunity where others see mess.

Security is the latest clean-up target in that playbook, making the push the newest example of ServiceNow applying its consolidation logic to a crowded category as it defends its position in the AI race.

1 source
Diginomica
ServiceNow & ITSM

The 320-day line that has never been wrong looms over NOW

Yahoo Finance flags a technical trendline with a historically perfect record, and the stock is stalling right at it.

Yahoo Finance flags a technical trendline with a historically perfect record, and the stock is stalling right at it.

Yahoo Finance notes that ServiceNow (NYSE: NOW) rallied from its April lows, but the recovery appears to be running out of steam near the declining 320-day moving average. That trendline has historically been a reliable bear signal when tested, which is why the article calls this the bear signal that has never been wrong.

With price action pinning itself against the descending long-term average, the piece frames the coming days as a make-or-break test of whether the post-April recovery extends or gives back its gains.

1 source
Yahoo Finance
Every edition

The wire publishes daily and keeps an archive of every edition, listed by date, newest first. If a summary is wrong, we correct or remove it.

Back to the archive Latest edition