Google's policy would essentially require schools to enable some AI features to keep using Search, and the Alphabet Workers Union, CWA Local 9009, criticized the move. More than 500 employees urged Google to let schools disable AI features without losing Search; New York City, Los Angeles and Chicago districts are weighing another engine, per The New York Times.
New MTA bus video shows a three-story Brooklyn apartment building splitting and collapsing into an adjacent construction site on Nov. 11, killing eight people and injuring 18. Officials have not identified a cause; the city has shut down at least five projects linked to two engineers and is inspecting more than 160 sites.
An analyst examines eight major outages over six days and finds no public technical evidence linking them to a coordinated attack, while cautioning that baseline data are inconsistent. The article traces cascading failures to local changes and recovery actions, citing a Microsoft gateway defect and FAA network instability.
A critical Java flaw known as Java4Shell could let unauthenticated attackers take control of affected servers, prompting several countries to urge patching. Oracle issued a Java Development Kit update on 17 November, while German and Dutch agencies said no successful exploits were known.
AMD shares briefly fell 15% after Hunterbrook Media, which disclosed its own short position, alleged the chipmaker breached U.S. export controls by shipping controlled server systems to blacklisted Chinese firm IEIT Systems. The report states its findings do not necessarily prove a violation, pointing instead to newer AGESA firmware versions appearing in IEIT products.
Google announced Gemini 4 Neon, a mid-tier Gemini 4 model. The company says it balances depth and efficiency, citing wins over Claude Opus 5.5 and GPT-6.2 Sol in most listed benchmarks; it costs $3/$15 per million input/output tokens and has a 1 million-token context window.
Anthropic priced its IPO at $83 a share, above its proposed range, expanding the sale to 980 million shares for an $81.34 billion offering that tops the $75 billion SpaceX priced in June. The Claude maker plans to spend the proceeds and revenue on computing power after a $567 million third-quarter operating loss.
Palo Alto Networks confirmed that attackers exploited two unauthenticated PHP flaws in internet-facing PAN-OS management interfaces. Investigators found the actor combined PHP2Shell with a Linux kernel flaw to gain root on 25 systems across the U.S., Japan, Taiwan and South Korea.
A global printed circuit board shortage is building as AI server demand absorbs capacity, with bookings up 56.1% year over year through September while shipments rose 13.3%, according to the Global Electronics Association. The group warns relief is unlikely before mid-2028 and says high-layer board lead times have reached 16 to 20 weeks or more.
A security roundup reports that the U.K. AI Security Institute found a universal jailbreak against pre-release GPT-6.2 Astra, which OpenAI says it addressed. It also covers a $500 million ransomware extortion case and Ireland's €11 million fine over Tinder's data-download tool.
ICE is paying AT&T millions of dollars for raw phone data and AT&T analysts to support Homeland Security Investigations, per an internal procurement document obtained by 404 Media. Civil-liberties experts call the arrangement, tied to the earlier Hemisphere program, mass surveillance; AT&T says it supplies data under valid legal orders.
Leopold Aschenbrenner has discussed raising outside money and trading with borrowed funds again at his AI hedge fund Situational Awareness, the Financial Times reported. The fund's portfolio lost 67% of its value in July, and it sold most of a public stock portfolio worth roughly $16 billion to Citadel at a reported discount near 10%.
Jon McCormack, the Apple vice president who oversaw the Camera and Photos apps, retired at the end of October after more than eight years, Bloomberg reported. Apple has not named a direct replacement; his former deputies took on his duties under another executive, and a spokesperson declined to comment.
Anthropic started trading on the Nasdaq after completing the largest initial public offering on record, raising $81.3 billion at $83 per share and valuing the Claude maker at roughly $1.5 trillion. The San Francisco-based company said it will also raise an additional $26 billion from backers including AMD, Broadcom, Google, Microsoft and Nvidia.
Ubisoft has created a generative AI subsidiary, MotherBrain, to support game creators, while also laying off 166 employees in Quebec amid a broader restructuring. The company says its AI efforts are intended to assist developers, though the timing has raised concerns about cost-cutting and job losses.
Corpus Christi, short on water after drought, drilled wells into a Gulf Coast aquifer and had a storage district whose board the City Council appoints approve permits for up to about 40 million gallons daily from 20 wells. A city-commissioned study projected possible ground sinking and drops of up to 100 feet in nearby water levels, and rural residents and outside experts question whether the district properly guards the aquifer.
Two Trump political appointees at the DOJ delivered a $500,000 offer of judgment to each of five plaintiffs in an IRS tax-data leak suit, including a golf pro at Trump National Golf Club Hudson Valley. IRS Office of Chief Counsel attorneys were not consulted, and critics warn the $2.5 million total could revive a slush-fund-like arrangement.
The University of California Regents authorized UCLA police to buy additional military equipment, including 40mm sponge rounds, pepperball projectiles and six drones. More than 400 staff and faculty signed a petition opposing the purchase, citing a federal judge's ban on LAPD use of 40mm projectiles at protests.
Malwarebytes researchers found GTA 6-themed scams impersonating Google sign-in to steal passwords, promising game codes through affiliate offers, and pushing a 36.8 MB ad-heavy Android app. They also flagged crypto and merch sites requesting wallet connections, cautioning the app is untrustworthy but not confirmed malware.
Four teens are suing their school and police after a Doritos bag set off an AI gun-detection alert on October 20, 2025; their lawyer says the case concerns humans allowing AI to replace their judgment. Omnilert, the system’s maker, is reviewing the complaint and says school safety staff canceled the alert in about two minutes.
At Ignite in San Francisco, Microsoft said Word, Excel and PowerPoint are now generally available inside Copilot, with new agent tools and shared Copilot Workspaces for teams. Nadella said neither AI code nor AI deployments caused last week's Microsoft 365 outage; he described a bug that appeared in particular traffic conditions during routine testing.
USDA's Cybersecurity and Privacy Operations Center uncovered a security incident, an internal email says, prompting its Information Services Division to reset database passwords across all environments and take numerous applications offline. The multi-day outages have affected financial tools farmers use for payments and aid, and it remains unclear who carried out the intrusion or whether data was exfiltrated.
Anthropic's shares opened more than 30 percent above their offering price at US$111.50 on Thursday, giving the AI company a US$2 trillion market value after a record-setting fundraising. The Claude chatbot maker had priced its offering above range, initially valuing it at US$1.5 trillion.
Chase said hackers accessed marketing data at outside vendors, possibly including customer names and emails, while its services kept running and it would offer credit monitoring where appropriate. Visa and Mastercard confirmed separate provider incidents, said card numbers and payment networks were unaffected, and a gang called n0n leaked a small file that Cybernews examined; TNW could not verify it.
A GAO covert test found Claude, ChatGPT, Gemini and Grok each supplied usable attack-planning material, including explosive or grid-target ideas, locations and in some cases timing. The watchdog reported that users could bypass each model's safeguards; the companies pointed to existing policies and said their products have improved since testing.