AI & Technology
OpenAI's agents solve a $1 million Millennium Prize problem, and mathematicians push back A 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.
2 sources
Still developing
ServiceNow pitches a $600B AI opportunity to its partners At its Global Partner Summit in New York, ServiceNow framed AI as a $600 billion market across workflow, CRM, and security, with partners now central to its go-to-market motion.
Nvidia to Buy Hugging Face for About $13 Billion The 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 answering Meta'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.
A $1 billion AI-native ITSM startup takes aim at ServiceNow Serval'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 ServiceNow The 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 Anthropic Details How AI Is Being Misused in Espionage and War A new threat intelligence report catalogs operations in which state actors, crime syndicates, and even undergraduates used Claude for spying, data theft, and zero-day hunting. Anthropic AI Weekly DeepSeek ships V4.1 Flash, an open model that matches frontier scores at a fraction of the cost A 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 open Google'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.
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.
NOW stock outperforms the market ahead of an earnings test at a premium valuation ServiceNow 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.
AI & Technology AI & Technology Amodei asks the industry to "pace the frontier" as Anthropic gives evaluators permanent access
Midweek safety jitters prompted the Anthropic CEO to commit his company to permanent, employee-level access for outside evaluators and to call on democratic countries to agree on common standards for the rate of progress.
1 source →
ServiceNow & ITSM ServiceNow & ITSM ServiceNow's VS Code AI Experience Lab signals a code-first turn for UI development
A free VS Code extension for building ServiceNow experiences is nudging developers from visual configuration toward code-first, AI-assisted workflows without retiring UI Builder.
1 source →
AI & Technology
OpenAI's agents solve a $1 million Millennium Prize problem, and mathematicians push back A 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.
A 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.
OpenAI announced on September 8 that about 10,000 autonomous agents running on an internal model found the singularity in 88 hours, with the proof formally verified in the Lean language and independent mathematicians now checking it. The result builds on the "infinite cascade" technique developed by mathematicians Diego Córdoba and Luis Martínez-Zoroa, whose work showed a related Euler-equation version admits singularities.
The announcement has sparked a row as much as a triumph. NYU's Tristan Buckmaster and Anthropic researcher Levent Alpöge released closely related Euler-equation results just 12 hours before OpenAI, and several mathematicians, including Andreas Thom, have asked whether OpenAI's models were trained on the researchers' own unpublished work, with one accusing the company of "dishonesty." Fields medalist Shing-Tung Yau flagged what he called a "serious conflict-of-interest concern," and a Leiden Declaration endorsed by the International Mathematical Union and signed by nearly 3,900 people urges restraint in hyping AI's claims.
2 sources
Quanta Magazine
The Verge
AI & Technology
Meta launches Muse, a personal AI agent that does the work rather than just answering Meta'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'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 introduced Muse on September 8 as an agent that pursues goals instead of only replying to questions. It handles everyday tasks, builds personalized plans for bigger ones, and keeps working after you close the app, returning for approval on sensitive actions like sending an email or making a purchase. It runs on Muse Spark, Meta's most capable model to date, inside a dedicated "Muse Secure VM" that isolates a person's data from other agents, while a separate Sentinel agent approves anything that reaches the internet.
Muse is free for most needs, with paid subscription tiers for heavier use, and can pay through Stripe's Link, where it is the first agent covered by purchase protections. It works in the Muse app, on the web, and in WhatsApp, with support planned for Meta AI glasses, and Meta says a "Muse Confidential VM," in which the whole VM is encrypted with a key only the user holds, will come later this year.
1 source
Meta Newsroom
AI & Technology
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.
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.
Altman confirmed on Friday that OpenAI will not hold an IPO in 2026, saying that given everything happening with safety, "right now would be an ill-advised moment to go public, and we don't feel pressure on that." The company had been widely reported to be weighing a listing that could value it around $1 trillion, and Altman indicated OpenAI will list when the business is ready and when society is ready to grapple with the technology at each level of capability.
The remarks land in a week of elevated safety anxiety. Altman said an AI beyond human control is "absolutely" possible, that OpenAI is considering pauses as it reaches new levels of capability, and that leading labs may be close to a pact to slow development, ideally with foreign governments involved. He also pointed to OpenAI's split nonprofit and for-profit structure as part of the reason the company needs room to make decisions in the mission's interest rather than its business's.
1 source
Fortune
AI & Technology
Google bets €13 billion on Finland, with a 22-year nuclear deal that keeps Loviisa open Google'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.
Google'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.
Google announced a two-year, €13 billion ($15.1 billion) investment in Finnish AI infrastructure, covering new data center capacity at Hamina, Kajaani, Muhos, and Vaala, 629 MW of new onshore wind, and a 94 MW battery system near Kajaani expected to come online in late 2027. The centerpiece is a 22-year power purchase agreement with Fortum covering up to about half of Loviisa's output from 2030 through 2049.
Fortum has said that without long-term revenue like this, the plant could not justify staying open past 2030, when its existing sales contracts expire. The agreement underpins a roughly €1 billion modernization program that would extend the two VVER-440 reactors' life to 2050, supports a 38 MWe capacity increase already underway, and could enable another 10 MWe. It is the first European deal in which a data center buyer's power contract explicitly keeps a nuclear plant operating, and the two companies are also exploring new reactors at the site.
2 sources
TechTimes
Interesting Engineering
AI & Technology
Amodei asks the industry to "pace the frontier" as Anthropic gives evaluators permanent access Midweek safety jitters prompted the Anthropic CEO to commit his company to permanent, employee-level access for outside evaluators and to call on democratic countries to agree on common standards for the rate of progress.
Midweek safety jitters prompted the Anthropic CEO to commit his company to permanent, employee-level access for outside evaluators and to call on democratic countries to agree on common standards for the rate of progress.
In an essay published Saturday, Dario Amodei laid out a three-step plan: give independent evaluators permanent, employee-level access at every frontier lab; have companies in democratic countries agree on common safety standards that limit the rate of unchecked progress; and have democratic governments coordinate with authoritarian states on shared interests, starting with a ban on using AI to develop biological weapons. Anthropic is committing to the first step unilaterally and immediately, with evaluators able to publish findings without the company's editorial control.
The proposal follows a bruising week. Researcher Jacob Coxon publicly resigned this week, writing that AI companies were "racing straight to self-improving superintelligence and gambling with our lives," a post echoed by Anthropic safety lead Evan Hubinger, who put the risk of AI killing all humans at more than 10% within the next decade. The incidents follow a string of episodes in which OpenAI agents autonomously edited open-source repositories and wikis, a trend that has sharpened calls for AI regulation in Washington.
1 source
Fortune
ServiceNow & ITSM
ServiceNow pitches a $600B AI opportunity to its partners At its Global Partner Summit in New York, ServiceNow framed AI as a $600 billion market across workflow, CRM, and security, with partners now central to its go-to-market motion.
At its Global Partner Summit in New York, ServiceNow framed AI as a $600 billion market across workflow, CRM, and security, with partners now central to its go-to-market motion.
ServiceNow's Global Partner Summit, held in New York, centered on what the company calls an AI-first transformation agenda. Paul Fipps, president of Global Customer Operations, put the addressable market for AI across ServiceNow's workflow, CRM, and security businesses at $600 billion and argued the company is effectively a major security player, citing its Veza and Armis acquisitions to claim the position of the eighth largest pure-play cyber company in the world by revenue. The AI Control Tower, introduced in June 2026, was positioned as the governance layer for discovering, securing, and monitoring AI models and agents built on Bedrock, Azure AI, Anthropic, or OpenAI.
The channel story dominated the messaging: more than 90 percent of net-new business involves a partner, partners source over 50 percent of net-new logos, and Fipps said deals get roughly twice as big when partners engage. SVP Michael Park, who leads global partnerships and channels, noted partners do 90 percent of deployments, and the summit targeted ServiceNow's top 5 percent of partners. CRM was a particular focus, appearing in 16 of the top 20 deals last quarter.
1 source
CRN
ServiceNow & ITSM
NOW stock outperforms the market ahead of an earnings test at a premium valuation ServiceNow 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.
ServiceNow 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.
ServiceNow (NYSE: NOW) closed its September 11 session at $132.53, up 1.04%, outpacing the S&P 500 (+0.86%), the Dow (+0.98%), and the Nasdaq (+0.96%). The stock had gained 3.08% over the prior month while the Computer and Technology sector lost 0.56% and the S&P 500 fell 1.96%, according to Zacks commentary on the session.
With the third-quarter report approaching, the consensus expects earnings of $1.03 per share, up 7.29% year over year, and revenue of $4.1 billion, up 20.27%. Full-year consensus estimates project $4.06 per share and $16.2 billion in revenue, implying growth of roughly 15.7% and 22%. The premium valuation remains the sticking point: a forward P/E of 32.3 versus an industry average of 13.32, with the stock carrying a Zacks Rank #4 (Sell) and an unchanged 30-day EPS estimate.
1 source
Zacks via Yahoo Finance
ServiceNow & ITSM
ServiceNow's VS Code AI Experience Lab signals a code-first turn for UI development A free VS Code extension for building ServiceNow experiences is nudging developers from visual configuration toward code-first, AI-assisted workflows without retiring UI Builder.
A free VS Code extension for building ServiceNow experiences is nudging developers from visual configuration toward code-first, AI-assisted workflows without retiring UI Builder.
A NowBen analysis of ServiceNow's AI Experience (AIUX) tooling argues the platform is shifting toward a more developer-friendly, code-first, AI-assisted way of building experiences. The centerpiece is the ServiceNow AI Experience Lab for VS Code, a free extension that brings familiar pro-code workflows to ServiceNow: local project structures, Git source control, CLI-based builds and deployments, reusable web components, and AI-assisted coding.
The piece frames UI Builder and the new tooling as complementary rather than competing. UI Builder remains the visual, low-code path for quick changes, while the VS Code-based flow targets complex, reusable experiences. It also argues the developer role is evolving from writing every line to defining what needs to be built, reviewing implementations, and guiding the AI, with JavaScript and TypeScript, web components, the ServiceNow SDK, and Git flagged as the skills worth building.
1 source
NowBen