Beyond the announcement.

The US midterms strengthened candidates and measures favoring restrictions on data centers, increasing pressure on a central pillar of artificial intelligence expansion. An NBC News exit poll cited by The Verge found that 76 percent of voters opposed data centers in their communities. The outlet also counted 45 local measures, 33 of which restricted development; Ars Technica cited a different tally from Redbud Consulting, with 44 measures across 10 states and 38 passing.

The victories of Democrats who support moratoriums, including Gina Hinojosa in Texas, could delay new projects. According to Ars Technica, citing Bloomberg, about 13 gigawatts of capacity in Texas could be exposed to delays. TechCrunch reported that Democratic control of Congress could create room for nationwide limits, while Wired reported that opposition to data centers crossed party lines and shaped local and statewide races.

Sources: The Verge, Ars Technica, Wired, TechCrunch

Read the original story ↗

Sam Altman defended his comments about buying RAM, writing on X that he wants people to put “as much as they want” in their computers and has long argued for increasing global memory production. He also changed his name on the platform to “Ram Altman” and wrote that he wanted people to have “so much RAM” and that more should be built.

The remarks followed an interview on the Cheeky Pint podcast in which Altman said it was “quite terrible” for large amounts of the limited DRAM supply to sit in personal computers and be used only 5% of the time. The Verge reports that AI data centers are taking up the world’s RAM supply, contributing to shortages and higher prices. Intel expects the shortage to last through at least next year, while SK Hynix predicts it could continue until 2030.

Sources: The Verge

Read the original story ↗

Google acknowledged to investors that strong query growth in AI Mode is accompanied by lower monetization. According to Xataka, the format uses Gemini to answer with generated text instead of showing a list of links, reducing the space for ads and the reasons to click them. Google chief commercial officer Philipp Schindler said many AI Mode queries are incremental, but did not say how much less they earn.

According to Xataka, Alphabet posted $127.577 billion in revenue in the third quarter, with search contributing $63.350 billion. The company spent $51.355 billion on equipment and facilities during the period, more than the $42.290 billion generated by its business. A Pew Research Center study cited by Xataka found that 8% of visits ended in a click on a reference link when an AI summary was present, compared with 15% without the summary; Google challenged the study’s methodology.

Sources: Xataka

Read the original story ↗

Reuters published an analysis saying China’s AI bubble has already burst. The STAR50 index, which tracks leading AI and semiconductor companies in mainland China, fell 40% from its July peak, erasing $480 billion in market value; the CSI300 declined 20% over the same period. Bloomberg described the STAR50 as the worst-performing major technology index in the world this semester.

According to Bloomberg, global AI hardware sales and a rush of foreign investors into the sector during the previous quarter put further pressure on Chinese stocks. Reuters says there is no clear trigger, although high valuations, taxes and possible Western restrictions on Chinese components are among the cited factors. Goldman Sachs says AI stocks account for 42% of China’s market capitalization and 57% of trading on mainland and Hong Kong exchanges; even after the decline, the STAR50 still trades at about 100 times its earnings over the past 12 months.

Sources: Xataka

Read the original story ↗

Some graduates are being hired for roles dedicated to AI. Lotte Underwood, at public relations agency Fight or Flight, manages about 30 agents that conduct research, produce reports, prepare media interviews and generate images. Dillan Patel, at engineering software company Stride, develops internal agents and identifies repetitive tasks that could be automated. The BBC reports that these roles give graduates broader responsibilities, but also raise questions about how much time they have to learn the traditional skills of their fields.

ArvatoConnect has put its usual graduate intake on hold this year, but CEO Debra Maxwell says it plans to resume it next year with a greater focus on AI-related roles. PwC analyzed 2.4 million entry-level jobs in the US for its 2026 AI Jobs Barometer and said the entry-level positions most influenced by AI are seven times more likely to require skills traditionally associated with senior professionals; those roles grew 35% since 2019, while other entry-level roles shrank by 10%.

Sources: BBC

Read the original story ↗

Some AI data centers are turning to jet engines removed from retired aircraft and converted to burn natural gas, according to The New York Times, cited by TechCrunch. The alternative can be brought online faster: Rystad Energy analyst Audun Martinsen told the Times that the units can get a data center running in 24 months, while connecting 100 megawatts to the grid in Northern Virginia takes about seven years, according to the International Energy Agency.

xAI has already installed eight such turbines in Mississippi to help power Colossus 2, the Times reported, and OpenAI used gas turbines at its first Stargate site in Texas in 2025. Crusoe ordered 13 units but walked away from a $1.25 billion deal for 29 turbines. The converted engines are less efficient and more polluting than conventional turbines, potentially making them a short-term solution.

Sources: TechCrunch

Read the original story ↗

OpenAI says the volume of searches made through ChatGPT has reached 20% of Google’s search volume. Adam Fry, ChatGPT’s consumer product lead, made the statement at a press event and said the company counts any query that results in a web search as a search, excluding tasks such as defining a term or giving a simple how-to explanation. The Verge noted that Fry did not identify the source of Google’s figure.

According to The Verge, Satya Nadella said during Google’s 2023 antitrust trial that Bing had about 3% of the search market after more than a decade of operation. Fry also said ads had not harmed the product or its usage and that most people did not seem to mind them. OpenAI announced last month that it would begin showing images of sponsored products and services in ChatGPT, rather than only a company name, logo and link.

Sources: The Verge, Tecnoblog

Read the original story ↗

On the MIT Technology Review Brasil podcast, Rafael Coimbra, Carlos Aros and Alexandre Roldão examine the progress of technologies that can capture brain signals and convert them into images or other understandable information. According to the discussion, these tools could expand communication for people with motor and speech limitations.

The episode also addresses mental privacy, consent and the limits of access to individual thoughts. The participants further discuss the possibility of reconstructing memories and creating digital replicas from brain records, as well as the risk that algorithms could treat fragments of human experience as complete representations.

Sources: MIT Technology Review Brasil

Read the original story ↗

Google announced that Gemini, NotebookLM, Gemini in Classroom, and Google Search will be disabled by default for K-12 student accounts when the changes roll out, unless school administrators enable them. Chris Phillips, Google’s vice president of education, said “Search and AI are deeply integrated”; as a result, schools that disable the AI features entirely will not have access to Search. Google says schools will be able to control the services, but Search cannot be separated from its AI features.

The decision drew criticism from Senator Bernie Sanders, former Federal Trade Commission (FTC) chair Lina Khan, New York City Mayor Zohran Mamdani, and other leaders, who accused Google of pressuring schools and acting as a monopoly. Google also announced Guided Learning, a feature intended to guide students through solving problems rather than simply provide answers; The Verge says it will begin testing this month in select US schools and will be integrated into Search’s AI Overviews.

Sources: TechCrunch, The Verge, Ars Technica

Read the original story ↗

Binance says the $6.3bn attack was prepared from June and involved a multi-stage campaign. According to the exchange, one engineer downloaded a tampered tool and, in September, another employee received a fake job offer, took part in interviews with supposed company representatives and installed malware disguised as a programming assignment. The attackers used the compromised computer and an internal artificial intelligence tool to reach the systems of two other engineers and begin unauthorised transfers on 29 October.

Binance suspended withdrawals globally after detecting the incident and says it has replenished 98% of the affected assets. The exchange also says about $2.85bn was frozen or intercepted with help from partners, according to Tecnoblog. The FBI attributed the attack to a group linked to North Korea, which rejected the allegation. Binance says it will increase investment in artificial-intelligence detection, hire external experts and work to establish a security research fund.

Sources: BBC, Tecnoblog, g1

Read the original story ↗

Apple updated its App Store rules to require minimum 18+ ratings for dating apps and AI companions that offer romantic interactions or sexual content. Developers must also disclose whether an app allows users to share their physical location with another person or group. In Brazil, the minimum rating will be A18 for dating apps and A16 for those sharing real-time location, although family services may request an official rating from the Ministry of Justice and Public Security (MJSP).

The new fields will be available in App Store Connect on November 16. According to Tecnoblog, all apps must submit this information by January 12, 2027; Apple says the change is part of a broader review of the App Store’s age-rating system.

Sources: TechCrunch, Tecnoblog

Read the original story ↗

In an update to its filing for a possible public offering, Anthropic reported $17 billion in quarterly revenue, up from $11.5 billion in the previous quarter, according to TechCrunch. The company posted a $567 million operating loss and a $23.3 billion net loss, much of it from an accounting charge related to convertible notes, according to Xataka. Revenue grew 47%, but compute consumed 58% of it, up from 48% in the previous quarter.

Anthropic has accumulated about $895.4 billion in spending commitments, mostly for compute and infrastructure. According to TechCrunch, its compute bill reached $10 billion for the quarter, including $5 billion in short-term commitments that cost significantly more than comparable multiyear agreements. The company attributes the pressure to demand, which forced it to secure additional capacity, while it works to expand its own infrastructure. The prospectus still does not disclose a price range or the number of IPO shares.

Sources: Xataka, TechCrunch

Read the original story ↗

OpenAI has started using the term “compute campus” for its data centers while announcing the Grid Protectors Program, which will begin at an upcoming site in Effingham County, Georgia. The initiative promises that the company will pay for the costs it creates and help communities keep the power on, especially during emergencies and extreme weather. In the announcement, “compute campus” appears far more often than “data center,” and OpenAI says its new facilities differ from those built a few years ago because they allow flexibility in power demand.

The Verge says the terminology appears to be new for OpenAI and reports that Axios had previously reported plans to rename the sites “AI compute campuses” as part of a policy and public-relations effort. Holly Elmore of PauseAI US and Abre’ Conner of the NAACP told The Verge that changing the name would not reduce environmental impacts or harm to local health, the environment and resources.

Sources: The Verge

Read the original story ↗

Wired interviewed Zohran Mamdani, New York City’s mayor, about what he has learned during his first 10 months in office and about issues in the midterm elections, including data centers, AI in schools, social media as a governance tool and the future of the Democratic Party.

The magazine describes Mamdani as an influential figure on the American left and says his approval among city voters stands at 69 percent. The article also revisits his political rise, driven by a message focused on affordability and the working class and by his effective use of social media.

Sources: Wired

Read the original story ↗

Elon Musk says artificial intelligence infrastructure companies may be close to a financial crisis. Jensen Huang, Nvidia’s CEO, disagrees and says demand for processing will continue to support the growth of so-called neoclouds, companies that rent large pools of GPUs and other specialized resources to developers, startups and technology companies.

According to Synergy Research Group, the sector’s global revenue exceeded US$25 billion in 2025 and could approach US$400 billion by 2031. The model requires billions of dollars in GPUs, data centers, energy and networks, while equipment becomes outdated quickly. At the same time, companies such as Microsoft and Meta are among neocloud customers. Nvidia is also investing in the sector, including US$2 billion in CoreWeave and US$2 billion in Nebius, announced in 2026, placing Huang at the center of the expansion he supports.

Sources: MIT Technology Review Brasil

Read the original story ↗

According to MIT Technology Review, an executive order signed by Trump at the end of September directed government agencies to call artificial intelligence “Super Intelligence,” or “SI.” The stated rationale is that current systems amplify human ingenuity rather than merely imitate intelligence. The term has already appeared in government solicitations and was silently added to many official websites, while some agencies display notices that both expressions may appear during the transition.

The publication argues that the change could go beyond vocabulary and become a political loyalty test, with use of the new term signaling support for the administration. The rollout has had lapses, however: the Pentagon used the old terminology for a new project, a Federal Trade Commission release used “superintelligence” in a different form, and recent Health and Human Services releases showed no sign of the change. Government employees and even Defense Secretary Pete Hegseth also continued to use “AI.”

Sources: MIT Technology Review

Read the original story ↗

Apple indicated during its earnings call that it may pursue new opportunities in third-party apps and partnerships to contribute to Services revenue growth. Kevan Parekh, the company’s chief financial officer, did not detail what those opportunities might be. Apple also expects Services growth of 8% to 9% in the December quarter, according to its forecast presented during the call.

The move comes as regulatory changes and court decisions pressure App Store commission revenue, alongside a slowdown in mobile gaming. According to Reuters, investors are considering whether Apple could distribute third-party AI agents through the App Store and take a share of the revenue they generate, but the company has not confirmed that possibility. TechCrunch contacted Apple for more information.

Sources: TechCrunch

Read the original story ↗

Trajectory, a startup founded about a year ago by three Stanford classmates, announced a $45 million Series A led by Sequoia, with participation from Nvidia, Bessemer Venture Partners, and Bain Capital Ventures. According to TechCrunch, the 18-person company is now valued at $300 million and has made its AI model post-training platform publicly available.

The platform connects open-source models to companies’ applications, collects feedback on their performance, and uses it to retrain them for specialized tasks in an approach known as continual learning. Co-founder Ronak Malde says this could let lawyers, doctors, and other specialists improve models almost automatically. Harvey, Rogo, Decagon, Mercor, and Nvidia are among its customers. Co-founder Michael Elabd says human involvement makes the approach safer than autonomous self-improvement, while companies including Thinking Machines Lab, Applied Compute, and Prime Intellect are also competing in the post-training market.

Sources: TechCrunch

Read the original story ↗

The New York Times sued Brave in federal court in Manhattan, accusing the browser and search company of copying copyrighted articles without permission and distributing them to customers including Anthropic, Perplexity, Mistral, xAI, and Cohere. The lawsuit says the company’s index contains more than 30 billion web pages and is commercialized through the Brave Search API, including to provide text and other data to language models. It also challenges the Paywall Blocker feature and filters that can bypass subscriber barriers, as well as Brave’s ad blocking for publishers.

In a separate lawsuit, New York Daily News, Chicago Tribune, and Orlando Sentinel made similar allegations, saying Brave offers paid access to real-time content from its index. Graham James, a Times spokesperson, told The Verge that the company ignored requests to stop the conduct. Brave did not immediately respond to requests for comment; Perplexity’s Jesse Dwyer criticized the lawsuit.

Sources: The Verge, TechCrunch, Tecnoblog

Read the original story ↗

Google says AI Mode monetizes worse than traditional search. Philipp Schindler, the company’s chief business officer, said the strong growth in queries through the format affected monetization; as a result, Google is expanding ads in AI Overviews and AI Mode to more users and queries, including international markets and multi-turn conversations in the United States. Schindler said the company’s ad experiments produced promising results.

Xataka argues that AI has made search better at answering questions directly, reducing the need to open other results. That may reduce ad exposure and value, creating a tension between offering more useful answers and preserving search revenue.

Sources: Xataka

Read the original story ↗