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

Edition, Tue, 15 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

At 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.

1 source
Lead
this edition
Still developing
OpenAI's agents solve a $1 million Millennium Prize problem, and mathematicians push backA swarm of 10,000 OpenAI agents produced a Lean-checked proof that 3D Navier-Stokes flow can develop a singularity in finite time, the first AI claim to one of math's great unsolved problems.Quanta MagazineThe Verge
Money rotated from AI hardware into enterprise software after Anthropic CEO Dario Amodei urged frontier AI labs to slow the pace of model capability development.
Nvidia to Buy Hugging Face for About $13 BillionThe GPU giant has signed a definitive agreement to fold the leading open-model hub into its own portfolio.AI Agent Store
Meta launches Muse, a personal AI agent that does the work rather than just answeringMeta's most ambitious consumer AI bet yet is a private agent that sends email, books travel, fills forms, and buys things on your behalf, rolling out this week in the US on iOS, Android, and WhatsApp.Meta Newsroom
A $1 billion AI-native ITSM startup takes aim at ServiceNowServal's Catalyst agent, now generally available, writes code-based workflow automations, and the company says it resolves employee requests rather than just tracking tickets.ComputerWorld
IBM and Marafiq move Saudi service delivery onto ServiceNowThe Kingdom's integrated utilities group Marafiq is migrating its service management environment to ServiceNow's cloud platform, with IBM Consulting leading the full implementation.IBM MEA Newsroom
Facing years without profitability and competition from cheaper open-weight Chinese models, Anthropic and OpenAI are using the existential risk narrative to pressure the U.S. government for regulatory protection that could make them "too big to fail."
DeepSeek ships V4.1 Flash, an open model that matches frontier scores at a fraction of the costA 552B mixture-of-experts release that undercuts GPT-6 Astra's price per run by more than 70x.The Neuron
Google bets €13 billion on Finland, with a 22-year nuclear deal that keeps Loviisa openGoogle's largest single investment in Europe is anchored by a power purchase agreement that gives the Loviisa nuclear plant the revenue certainty to keep operating through 2050, a first for the continent.TechTimesInteresting Engineering
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
NOW stock outperforms the market ahead of an earnings test at a premium valuationServiceNow closed September 11 at $132.53, up 1.04%, beating the S&P 500's 0.86% gain as investors weigh roughly 20% revenue growth estimates against a rich multiple.Zacks via Yahoo Finance
Apple's First Foldable iPhone Costs $1,999The iPhone Duo opens to a 7.6-inch display, ditches Face ID, and hits stores October 23.Apple NewsroomMacRumors
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
Anthropic CEO Dario Amodei published an essay titled "We Must Pace the Frontier," urging the AI industry to slow its pace of development, and got an unusual chorus of support from rivals Sam Altman, Demis Hassabis, and Elon Musk before President Trump dismissed it as a hoax.

AI & Technology

AI & Technology
The French energy giant and Paris-based AI developer have signed a three-year partnership to develop frontier AI models for reservoir exploration and engineering, drawing on a century of geological data.
1 source
AI & Technology
A Mozilla report released today finds that the performance gap between U.S. frontier models and China's best open-weight models has shrunk to just 4.4 months, explaining why companies are rapidly shifting to far cheaper open alternatives.
1 source

ServiceNow & ITSM

ServiceNow & ITSM
Needham, BTIG, and BMO Capital all raised their targets on Monday following strong Q2 results.
1 source
ServiceNow & ITSM
The AI product line moved beyond $1B in ACV during Q2, with agentic deployments up ninefold in nine months.
2 sources
AI & Technology

Salesforce and Nvidia launch Koa, an enterprise reasoning model that challenges frontier labs

At 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.

At 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.

Koa represents a significant shift in how enterprises approach AI. Rather than routing all reasoning through expensive frontier models like Claude or GPT, Salesforce customers can now use Koa through the Agentforce platform for multi-step tasks. The model was post-trained on synthetic data that simulated customer service environments, without using any actual customer data. Salesforce and Nvidia argue Koa is more cost-effective, using fewer tokens than frontier alternatives while keeping customer data within Salesforce's infrastructure. The move underscores a growing divergence between what enterprise customers need — task-specific reasoning — and what frontier labs are building — general-purpose superintelligence.

1 source
TechCrunch
AI & Technology

Anthropic and OpenAI seek U.S. government backing to become too big to fail

Facing years without profitability and competition from cheaper open-weight Chinese models, Anthropic and OpenAI are using the existential risk narrative to pressure the U.S. government for regulatory protection that could make them "too big to fail."

Facing years without profitability and competition from cheaper open-weight Chinese models, Anthropic and OpenAI are using the existential risk narrative to pressure the U.S. government for regulatory protection that could make them "too big to fail."

Despite hundreds of billions in investment, both companies remain far from profitability and heavily dependent on capital inflows. The "existential threat" framing, amplified by Anthropic researchers Jacob Coxon and Evan Hubinger warning that AI could kill all humans within a decade, gives these firms a compelling argument for government intervention. The Register notes this strategy serves multiple purposes: it resets investor expectations by buying time, leverages U.S. national security concerns about Chinese models like DeepSeek, and could transform these startups into entities deemed too critical to let fail. The approach has drawn sharp criticism from those who see it as "safety-washing," including Nvidia's Jensen Huang who suggested the fear narrative "creates demand."

1 source
The Register
AI & Technology

TotalEnergies and Mistral invest €100 million in oil exploration AI

The French energy giant and Paris-based AI developer have signed a three-year partnership to develop frontier AI models for reservoir exploration and engineering, drawing on a century of geological data.

The French energy giant and Paris-based AI developer have signed a three-year partnership to develop frontier AI models for reservoir exploration and engineering, drawing on a century of geological data.

The program involves a joint scientific laboratory staffed by TotalEnergies' subsurface specialists and Mistral's machine learning researchers, running approximately 10 petaflops of computing power on TotalEnergies' proprietary subsurface data. The goal is to help geoscientists determine the location and size of oil and gas reservoirs and develop optimal extraction strategies over time. The partnership also explores "Agentic AI" — autonomous models that could not only identify promising formations but work through follow-on engineering questions. The deal emphasizes building a European AI ecosystem to reduce dependency on U.S. cloud providers, a theme also seen in Mistral's partnerships with EDF for nuclear engineering and Veolia for water management.

1 source
Energy Digital
AI & Technology

Open AI models close the gap to frontier labs to just four months

A Mozilla report released today finds that the performance gap between U.S. frontier models and China's best open-weight models has shrunk to just 4.4 months, explaining why companies are rapidly shifting to far cheaper open alternatives.

A Mozilla report released today finds that the performance gap between U.S. frontier models and China's best open-weight models has shrunk to just 4.4 months, explaining why companies are rapidly shifting to far cheaper open alternatives.

According to the State of Open Source AI report, Moonshot AI's Kimi K3 achieves a composite AI performance score within three points of Anthropic's Fable 5, while costing only 30 percent as much. The finding reveals a narrow band of workloads where paying for closed models is justified: expert professional work, high-intensity retrieval, and long-context tasks. Mozilla CTO Raffi Krikorian noted that "closed earns its premium in a few places" but argued that most organizations should use open models as the default for the majority of their work. The report highlights a fundamental tension: U.S. labs charge a premium for compliance packaging, support, and accountability, while open models allow anyone to download and run the main components on their own infrastructure.

1 source
Ars Technica
AI & Technology

The great AI slowdown debate: Amodei, Altman, and Musk call for a pause — Trump pushes back

Anthropic CEO Dario Amodei published an essay titled "We Must Pace the Frontier," urging the AI industry to slow its pace of development, and got an unusual chorus of support from rivals Sam Altman, Demis Hassabis, and Elon Musk before President Trump dismissed it as a hoax.

Anthropic CEO Dario Amodei published an essay titled "We Must Pace the Frontier," urging the AI industry to slow its pace of development, and got an unusual chorus of support from rivals Sam Altman, Demis Hassabis, and Elon Musk before President Trump dismissed it as a hoax.

Dario Amodei outlined three specific steps: embedding third-party evaluators inside AI labs, coordinating safety standards among frontier companies in democratic countries, and pursuing global coordination on pacing. Sam Altman, Demis Hassabis, and Kamala Harris all voiced support. Elon Musk simply wrote "Dario is right." The proposal sparked a fierce backlash from the Trump administration, which called the fears a "hoax" and "scam," and from Nvidia CEO Jensen Huang, who told Trump on stage at the All-In Summit "we are not going to let an AI slowdown happen." Underlying the debate is a fear of recursive self-improvement, a milestone where AI models could train new versions of themselves without human involvement, which Anthropic suggests could arrive as soon as early 2027.

3 sources
The Verge
The Verge
BBC News
ServiceNow & ITSM

ServiceNow surges 5-7% as AI slowdown flips the software trade

Money rotated from AI hardware into enterprise software after Anthropic CEO Dario Amodei urged frontier AI labs to slow the pace of model capability development.

Money rotated from AI hardware into enterprise software after Anthropic CEO Dario Amodei urged frontier AI labs to slow the pace of model capability development.

ServiceNow (NYSE: NOW) jumped roughly 6.4% to $140.98 in Monday regular trading as investors shifted toward enterprise software and unloaded AI infrastructure names. The Philadelphia Semiconductor Index sank nearly 6% in the same session, while ServiceNow and Adobe each gained more than 3%. The catalyst was a weekend essay by Dario Amodei urging labs to slow the pace of model capability development. Reuters reported the divergence, and sell-side analysts are now viewing ServiceNow's AI products as a revenue line rather than a threat to the company.

2 sources
Yahoo Finance
24/7 Wall St
ServiceNow & ITSM

Analysts lift ServiceNow price targets across the board

Needham, BTIG, and BMO Capital all raised their targets on Monday following strong Q2 results.

Needham, BTIG, and BMO Capital all raised their targets on Monday following strong Q2 results.

Needham lifted its price target from $115 to $155 (Buy), BTIG raised it from $150 to $170 (Buy), and BMO Capital moved it from $118 to $150 (Outperform). The consensus 12-month target sits around $145.71 with a Moderate Buy rating. The stock trades at a P/E ratio above 80, and analysts flagged committed remaining performance obligations (cRPO) as the "hard test" for the growth story.

1 source
AD HOC NEWS
ServiceNow & ITSM

ServiceNow AI products cross $1 billion in annual contract value

The AI product line moved beyond $1B in ACV during Q2, with agentic deployments up ninefold in nine months.

The AI product line moved beyond $1B in ACV during Q2, with agentic deployments up ninefold in nine months.

In its Q2 2026 results, ServiceNow reported that its AI products crossed the $1 billion annual contract value mark during the quarter, while agentic AI deployments multiplied ninefold in just nine months. Q2 subscription revenue was $3.877 billion, up 24.5% year over year. Current RPO reached $13.2 billion and total RPO hit $29.0 billion, both up 21% YoY. Non-GAAP EPS of $0.90 beat the consensus estimate of $0.86. The company said it is operating to a "Rule of 56" score, on track for "Rule of 60" by 2030, and that its AI Control Tower has more than 500 live customers within six months of launch.

2 sources
Yahoo Finance
AD HOC NEWS
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