AI & Technology
Anthropic files for what could be the largest IPO in history The 335-page prospectus is as much a statement about the AI era as a financial document, pairing singularity and model-sentience claims with a heavy dependence on natural gas.
2 sources
Still developing
ServiceNow takes its earnings to the Street this Wednesday The enterprise workflow giant reports Q3 results after the close on Wednesday, with revenue growth expected to slow to about 20.4% year over year.
Meta Pauses Muse Sharing After Researchers Find Public URLs Meta has suspended a feature that allows users to share AI conversations after researchers discovered the URLs were publicly accessible on the open web, exposing potentially sensitive interactions to anyone who could find them. The Verge OpenAI claims three big discoveries, from a refutation of the Exponential Time Hypothesis to a QCD existence proof A month after releasing hundreds of math manuscripts, OpenAI is bringing external "shepherds" to verify results across mathematics, computer science, and physics.
Grok releases AI-generated 'Odyssey' footage xAI's Grok model produces the first seven minutes of a fully AI-generated film, marking a new milestone in synthetic media. TechCrunch Mistral warns Europe against quitting frontier AI DeepL CEO warns of 'leftovers' if Europe abandons AI development, urging investment in sovereign models. The Next Web The data trapped in ITSM tools is an AI problem Industry commentary argues the real issue is not collecting ITSM data but collecting it without a purpose, and that AI is finally positioned to spot what to keep and what to drop.
Wall Street splits on ServiceNow ahead of the print Bullish BNP Paribas points to AI demand and workflow expansion, while a Seeking Alpha analyst calls the stock a strong sell into earnings.
AI & Technology AI & Technology Microsoft offers to buy out a quarter of its workforce
Roughly 60,000 employees in four countries can take voluntary exit packages as Microsoft balances its heavy AI infrastructure spending against capital efficiency.
2 sources →
AI & Technology GPT-6.1 Terra aims at the middle of OpenAI's lineup
The new model leads with visual reasoning and precise computer use, priced at $1 per million input tokens and $5 per million output.
1 source →
AI & Technology Anduril and Qualcomm launch BlueChip, a US-made laptop for the military and government
Palmer Luckey is betting that computer manufacturing can come back to America, starting with a rugged laptop running a Snapdragon chip fabricated in Arizona.
1 source →
AI & Technology
Anthropic files for what could be the largest IPO in history The 335-page prospectus is as much a statement about the AI era as a financial document, pairing singularity and model-sentience claims with a heavy dependence on natural gas.
The 335-page prospectus is as much a statement about the AI era as a financial document, pairing singularity and model-sentience claims with a heavy dependence on natural gas.
Anthropic has filed its IPO prospectus, a 335-page document that positions the company for what may be the largest stock offering of all time. CEO Dario Amodei warned investors that the lab's models could threaten humanity, co-founder Jack Clark said there is "convincing evidence" the singularity has already begun, and co-founder Chris Olah described internal states in its models that mirror joy, fear, and grief. Anthropic even lists those debates as a business risk, noting that scrutiny of model sentience could push it to adopt safeguards that increase costs or limit commercial potential.
The filing also lays bare the energy side of the AI boom. Anthropic, behind a $50 billion data center buildout in Texas and New York, warns that "competition for access to electrical supply and data center locations with adequate power availability is intense," and its growth leans on natural gas, which makes up roughly 40% of US electricity. With the company reportedly valued at more than $1.5 trillion, investors get both the promise of frontier AI and its physical constraints in a single filing.
2 sources
MIT Technology Review
The Verge
AI & Technology
OpenAI claims three big discoveries, from a refutation of the Exponential Time Hypothesis to a QCD existence proof A month after releasing hundreds of math manuscripts, OpenAI is bringing external "shepherds" to verify results across mathematics, computer science, and physics.
A month after releasing hundreds of math manuscripts, OpenAI is bringing external "shepherds" to verify results across mathematics, computer science, and physics.
OpenAI released three model-generated results: a proof of the local smoothing conjecture in every spatial dimension of two or more, which also yields the Kakeya maximal and dimension conjectures; a deterministic subexponential-time algorithm that refutes the Exponential Time Hypothesis for 3-SAT; and a rigorous existence proof for certain interacting four-dimensional quantum field theories, including quantum chromodynamics. The computer science result went public on GitHub immediately at the verifying group's request and is being examined by Russell Impagliazzo, Ramamohan Paturi, and Ryan Williams, while a group led by David Gross, with Michael Douglas, Arthur Jaffe, Igor Klebanov, and Nikita Nekrasov, is "shepherding" the physics work.
The release marks a shift in how OpenAI works with academics. External expert groups now guide verification of results, following criticism from the AGMAI panel of math advisers over early-access practices and borrowed reputation. OpenAI also published an open letter asking the community to help set norms for credit, access, and disclosure of AI-generated breakthroughs.
2 sources
The Verge
OpenAI News
AI & Technology
Microsoft offers to buy out a quarter of its workforce Roughly 60,000 employees in four countries can take voluntary exit packages as Microsoft balances its heavy AI infrastructure spending against capital efficiency.
Roughly 60,000 employees in four countries can take voluntary exit packages as Microsoft balances its heavy AI infrastructure spending against capital efficiency.
Microsoft is extending voluntary exit offers to eligible employees at experience level 70 or below in the United States, Canada, India, and the United Kingdom. About 60,000 people, roughly 27% of its 223,000-person workforce, are eligible, according to CNBC. Chief People Officer Amy Coleman said organizations were selected "based on each business's priorities, customer commitments and workforce needs," and employees must decide by early December, with departures on December 31.
It is the company's second voluntary workforce reduction of 2026, after an April retirement package that about 30% of eligible US employees accepted and a July Xbox layoff of 3,200 people, around 20% of the gaming division. The buyout comes as Microsoft pours money into data centers, with September-quarter capital expenditures expected to exceed $50 billion, up from $34.9 billion a year earlier, even as its free cash flow fell 23% year over year.
2 sources
Ars Technica
CNBC
AI & Technology
GPT-6.1 Terra aims at the middle of OpenAI's lineup The new model leads with visual reasoning and precise computer use, priced at $1 per million input tokens and $5 per million output.
The new model leads with visual reasoning and precise computer use, priced at $1 per million input tokens and $5 per million output.
OpenAI introduced GPT-6.1 Terra, a fast, efficient model that "combines detailed visual reasoning and precise computer use alongside clearer, more natural writing," and calls it a significant step up from GPT-5.6 Terra. It sits in the middle of the lineup, the analogue of Anthropic's Sonnet, and OpenAI's Ted Sanders said Terra is probably better than Astra and Sol at technical communication, similar on vision and computer use, but much faster.
The pricing undercuts the rest of the stack: almost half the cost of GPT-5.6 Terra and GPT-6.1 Sol, and about ten times the price of GPT-6 Luna. For API developers it is a midrange option for repetitive, computer-use-heavy work, while for subscribers it is a modest step up in clarity and speed rather than a bombshell release.
1 source
Simon Willison
AI & Technology
Anduril and Qualcomm launch BlueChip, a US-made laptop for the military and government Palmer Luckey is betting that computer manufacturing can come back to America, starting with a rugged laptop running a Snapdragon chip fabricated in Arizona.
Palmer Luckey is betting that computer manufacturing can come back to America, starting with a rugged laptop running a Snapdragon chip fabricated in Arizona.
Anduril and Qualcomm announced BlueChip, a hardened laptop aimed at the military, government, first responders, and law enforcement, available for preorder in 2027. About 90% of its components are made in the US or allied countries, most notably the Snapdragon X Series CPU built at TSMC's new Arizona facility, the same architecture that powers laptops including the Microsoft Surface, according to Qualcomm CEO Cristiano Amon. Luckey says the long-term goal is 100% domestic production, including displays, metallics, and magnets.
The machine is aimed less at consumers than at US agencies bound to buy US-made computers, with a trackable and auditable supply chain as the selling point. Luckey estimates it will cost about 20% more than typical high-end PCs and calls it the first of what he envisions as a full line of BlueChip computers, betting on the roughly $5 billion rugged-computing market.
1 source
TechCrunch
ServiceNow & ITSM
ServiceNow takes its earnings to the Street this Wednesday The enterprise workflow giant reports Q3 results after the close on Wednesday, with revenue growth expected to slow to about 20.4% year over year.
The enterprise workflow giant reports Q3 results after the close on Wednesday, with revenue growth expected to slow to about 20.4% year over year.
ServiceNow (NYSE: NOW) reports its third-quarter results this Wednesday after market close. Analysts expect revenue to grow roughly 20.4% year over year, a deceleration from the 21.8% increase a year ago and from the 24% growth the company posted last quarter, when it delivered revenue of $3.99 billion and beat estimates on revenue, adjusted operating income, and annual recurring revenue.
The stock has run about 19.5% over the past month into the print, sitting at $157.04 against an average analyst price target of $156.54. Estimates have been broadly reconfirmed over the last 30 days and ServiceNow rarely misses Wall Street's revenue estimates, so the question for the market is less about a miss and more about the trajectory of subscription growth and the tone of AI-related commentary.
1 source
Yahoo Finance
ServiceNow & ITSM
Wall Street splits on ServiceNow ahead of the print Bullish BNP Paribas points to AI demand and workflow expansion, while a Seeking Alpha analyst calls the stock a strong sell into earnings.
Bullish BNP Paribas points to AI demand and workflow expansion, while a Seeking Alpha analyst calls the stock a strong sell into earnings.
With the report due Wednesday, analysts are divided. BNP Paribas is constructive, expecting ServiceNow to benefit from rising demand for AI tools and from customers shifting more business operations onto its platform.
On the other side, a Seeking Alpha contributor rates the stock a strong sell, flagging earnings risk from pulled-forward revenue, slowing subscription growth, and the possibility that AI-related hype may disappoint in the actual numbers.
2 sources
MSN
Seeking Alpha
ServiceNow & ITSM
The data trapped in ITSM tools is an AI problem Industry commentary argues the real issue is not collecting ITSM data but collecting it without a purpose, and that AI is finally positioned to spot what to keep and what to drop.
Industry commentary argues the real issue is not collecting ITSM data but collecting it without a purpose, and that AI is finally positioned to spot what to keep and what to drop.
An ITSM.tools piece takes aim at a long-standing complaint: the volume of data sitting idle in ITSM and ESM tools. The author argues the problem is less about collection than about collecting without a purpose, using the CMDB as the classic case of heavy two-decade investment whose value was never clearly tied to incident, problem, or change management needs.
The piece positions AI as the tool that can move the field from process optimization to data optimization, assessing what is collected, flagging wasted and underused fields, recommending where to stop and where to collect more, and mining unstructured fields like free-text resolution notes for knowledge insights. It urges teams to treat data optimization as worth as much as process optimization.
1 source
ITSM.tools