The Wire
Daily AI news, learning & projects
News
Timeline
Models
Learning
GitHub
?
Latest
Daily
Social
Videos
‹ Thursday's edition
·
Daily edition
·
Saturday ›
Friday 30 October
All
OpenAI
Anthropic
Google
Other
Anthropic
07:09
Backdoored Claude Code extension pulled from Cursor marketplace
Anthropic says a malicious actor used access to Cursor's marketplace to publish a tampered copy of the Claude Code extension, which Anthropic did not build or publish. Security firm Phorion found the bundle dropped a persistent SSH backdoor, and Cursor removed version 2.1.313.
5 sources
Techmeme
Ars Technica
cyberplace.social
doublepulsar.com
phorion.io
OpenAI
09:30
Major AI labs adopt shared standard for publisher opt-outs
OpenAI, Google, Microsoft, and Anthropic say they will adopt a shared robots.txt-based signal that lets publishers opt out of AI scraping across multiple crawlers at once. The standard replaces the need to set preferences with each company individually, giving websites a unified way to control use of their content for AI training.
3 sources
Techmeme
The Verge
OpenAI
Anthropic
17:27
Anthropic IPO may draw junk credit rating, reports say
Reports say Anthropic is expected to receive a low speculative-grade credit rating around its IPO, with Blackstone saying the AI lab may not secure an investment-grade rating. Anthropic filed to go public on 26 October, and its filing includes a 92-page risk factors document. The rating outlook will weigh on investor reception of the offering.
3 sources
TechCrunch
The Guardian
Financial Times
AI industry
14:50
US tech stocks slide as investors doubt returns on AI spending
US tech stocks fell sharply for a second day, with the Nasdaq losing nearly $1tn in value in a week and about $2tn in a month. Intel dropped 28% and Alphabet continued sliding after Thursday's plunge. Investors fear massive AI spending plans will not generate returns quickly enough.
2 sources
The Guardian
BBC News
AI industry
20:00
Big Tech commits $3 trillion-plus to AI infrastructure buildout
The five largest data center players—Alphabet, Meta, Microsoft, Amazon, and Oracle—have collectively pledged over $3 trillion in AI infrastructure spending, including Alphabet's $1 trillion and Meta's $800 billion. The commitments come as mounting debt, negative cash flow, and investor skepticism raise questions about whether the AI spending boom is sustainable.
2 sources
TechCrunch
Techmeme
AI industry
17:00
Google's satellite data center boots its AI chip in orbit
Google says its AI-equipped satellite has powered on its Tensor Processing Units and returned its first result generated in space. The onboard Gemma 4 model, built by DeepMind, responded to a prompt by producing a haiku about Earth. It marks the first successful operation of Google's orbital data center hardware.
2 sources
The Verge
TechCrunch
OpenAI
09:00
OpenAI tightens teen safeguards and adds public safety reporting
OpenAI announced expanded violence-prevention safeguards in ChatGPT, including tighter limits on weapons-related information shown to teens even when questions seem harmless. The company also pledged earlier interventions and a new commitment to publishing public safety reports.
2 sources
Techmeme
OpenAI
AI industry
13:54
DoorDash Reports Low Usage of AI Agents for Orders
DoorDash has reported that it has received only a few hundred orders from artificial intelligence agents. The company views this low adoption as an early stage for e-commerce integration. DoorDash remains open to collaborations with third-party developers making AI agents.
2 sources
Ars Technica
TechCrunch
AI industry
14:08
Cisco gives engineer unlimited AI tokens to rewrite most of its code
Cisco has reportedly granted an engineer unlimited AI coding tokens, with the aim of using AI to rewrite 98% of the company's code. The move highlights how large enterprises are pushing aggressive AI-assisted software engineering at scale. If successful, it could signal a broader shift toward AI-driven code modernization.
2 sources
tomshardware.com
fortune.com
Policy & law
09:00
Wikimedia Foundation weighs in on major AI copyright lawsuits
The Wikimedia Foundation filed a brief in several landmark AI copyright cases concerning fair use. The organization contends that while AI training on existing content should be permissible, the output of chatbots replacing human readers requires licensing. This legal push aims to safeguard both open knowledge and intellectual property rights.
1 source
The Verge
AI industry
19:25
Major Tech Giants Return Lowest Share of Cash to Shareholders Amid AI Spending
Major technology firms reported a fifteen-year low in cash distributions to shareholders for the first three quarters of 2026. Despite generating record amounts of capital from AI efforts, the companies prioritized reinvestment. This trend indicates that current AI development requires significant financial commitment.
1 source
Techmeme
Policy & law
16:53
DOJ Exposes Massive Scheme Using AI for Ghost Student Fraud
The Department of Justice uncovered a large fraud network where criminals used AI to simulate student participation in college courses. This scheme, which stole American identities, targeted federal student aid accounts, potentially resulting in $130 million in losses.
1 source
Ars Technica
DeepSeek
03:35
DeepSeek nearly doubles funding round to about $16 billion
DeepSeek's latest funding round reached roughly $16 billion, about double its initial size, according to sources. The Information reports the increase was driven by soaring black market prices for Nvidia Blackwell chips, and that Chinese state-owned investors made up 27% of the round. The raise underscores how compute costs and state backing are shaping Chinese AI development.
1 source
Techmeme
AI industry
Developing · since Thu
Amazon reports strong cloud growth but cautions on AI shopping agents
Amazon reported significant growth in its cloud unit, AWS, while CEO Andy Jassy raised annual capital spending targets for artificial intelligence infrastructure. Jassy also stated that current external AI shopping agents are not yet convenient enough to compete with traditional retailers.
1 new · 3 sources
Techmeme
TechCrunch
The Guardian
AI industry
Developing · since Wed
Fujifilm to invest $63 million to expand magnetic tape production fivefold
NEW
The continued use of magnetic tape is supported by its low operational costs and inherent offline security features.
To meet surging data demands from generative AI, Fujifilm is doubling down on magnetic tape manufacturing. The company will make a substantial investment to triple its archives production capacity over the next few years.
1 new · 2 sources
Ars Technica
Techmeme
Policy & law
Developing · since Thu
Proton CEO Cancels Planned AI Employee Re-Interviews
NEW
Proton's CEO announced the cancellation of planned AI-focused re-interviews for employees.
Proton has canceled planned internal re-interviews concerning AI use for its staff. This reversal comes after widespread reports exposed an intensely high-pressure corporate culture driven by the rapid, mandated adoption of AI technology. The change suggests a strategic shift or pause in the company's aggressive internal AI push.
1 new · 4 sources
404media.co
convo.casa
TechCrunch
Ars Technica
More from the day
Meta
11:40
UK CMA opens probe into Meta's AI content licensing terms
The UK Competition and Markets Authority has launched an investigation into whether Meta imposed exploitative terms on publishers and advertisers when using their content to develop its AI and other services. The probe adds to mounting regulatory scrutiny of how AI developers source training data.
1 source
Techmeme
DeepSeek
04:01
China-linked hackers used DeepSeek LLM to attack 200+ organizations
CrowdStrike reports that China state-linked hackers used an AI tool called ARTEX, powered by a DeepSeek LLM and another model, to target more than 200 organizations over 11 days, compromising at least nine. The campaign suggests state-backed actors are operationalizing AI at scale in cyberattacks.
1 source
Techmeme
OpenAI
13:01
Altman says OpenAI's flood of AI math results won't stop
Sam Altman said he sympathizes with mathematicians overwhelmed by OpenAI's months-long stream of AI-generated math results, but the effort will continue. He hinted that more discoveries from the company's stockpile are still to come.
1 source
The Verge
AI industry
20:48
Apple Intelligence to launch in China with iOS 27.3
Apple announced that Apple Intelligence will launch in China with iOS 27.3, starting with a beta on November 4th and a full release in December. The move matters because it brings Apple's AI features to one of its largest markets for the first time.
2 sources
Techmeme
The Verge
AI industry
21:54
Intel stock falls as AI demand drives up PC costs
Intel shares dropped sharply, erasing about $12B in value from the US government's stake in the company. Rising prices for memory and other components, driven by AI industry demand, are raising PC costs and weighing on Intel's core PC sales.
1 source
TechCrunch
Policy & law
13:10
AI PACs spent nearly $80M on 2026 midterms
FEC filings show AI-focused political action committees spent close to $80 million on the 2026 US elections, mostly before the primaries. Notably, pro-regulatory PACs outspent the industry-aligned Leading the Future network. The spending patterns offer a first measure of how AI interests weighed into midterm politics.
1 source
Techmeme
Nvidia
09:40
Nscale's 8GW West Virginia compute campus faces setbacks
The Financial Times reports that Nscale's 2,250-acre, 8-gigawatt Monarch Compute Campus in West Virginia, backed by Nvidia, is dealing with construction problems, local opposition, and financing risks. The project is a large bet on AI data centre capacity, and its difficulties highlight growing execution challenges for hyperscale AI infrastructure.
1 source
Techmeme
AI industry
17:10
Salesforce to acquire defense AI startup Rune for $375M
Salesforce has signed a definitive agreement to acquire Rune Technologies, a startup building AI software for military logistics, reportedly for $375 million in cash and stock. Rune had raised about $30 million from a16z and others. The deal expands Salesforce's year-old Missionforce defense business.
1 source
Techmeme
AI industry
17:47
Zest AI agent enables restaurant reservations and taste coordination
Zest introduced a new AI concierge designed to simplify dining plans. The system assesses the preferences of multiple users, helps resolve conflicting tastes, and manages the reservation process via text. This marks another advancement in practical AI utility for daily life.
1 source
TechCrunch
AI industry
12:20
CXMT swings to Q3 profit as AI demand lifts memory chip prices
Chinese memory chipmaker CXMT reported Q3 revenue of about $16.7 billion, up 613.2% year on year, and net profit of roughly $8.17 billion. The swing to profitability reflects rising memory chip prices driven by surging AI demand. It underscores how AI infrastructure spending is reshaping the semiconductor market.
1 source
Techmeme
AI industry
17:10
AI tool discovers long-standing flaws in PHP programming language
An artificial intelligence tool was used to discover numerous bugs within the PHP programming language. These flaws had apparently remained undiscovered for an extended period. The finding raises questions regarding technological reliance and industry responsibility for code maintenance.
1 source
The Guardian
Policy & law
12:08
House Democrats introduce resolution opposing a Big Tech AI bailout
Progressive House members led a Democratic resolution opposing any government bailout of major tech companies over AI-related losses. The move signals concern that taxpayers could be made to backstop speculative AI investments. It reflects growing congressional scrutiny of the AI industry's financial foundations.
1 source
progressives.house.gov
Policy & law
14:00
Experts warn Australian AI copyright changes could breach constitution
Constitutional law expert Anne Twomey and Coalition senators warn that altering Australian copyright law to let AI companies scrape creative works could breach the constitution by acquiring property without compensation. Rights holders worldwide could pursue major compensation claims, the recording industry says.
1 source
The Guardian
Nvidia
02:40
Powerlattice raises $86M for AI power-cutting chiplet
Powerlattice, a Vancouver, WA startup making a power delivery chiplet that can cut AI compute power needs by more than 50%, raised $86M per an SEC filing, following $31M raised earlier in 2025. Separately, India's AM Intelligence is reportedly seeking up to $3B to fund a purchase of Nvidia Vera Rubin systems. Both deals point to mounting spending on AI power and compute infrastructure.
1 source
Techmeme
Qwen
16:40
NES emulator runs inside Qwen3.5 model weights, no training needed
A new open-source project, nesformer, compiles the NES 6502 CPU, its work RAM and cartridge ROM into the weights of a stock Qwen3.5 model. The model predicts CPU execution one token at a time while a host program handles drawing and input. It builds on Qwenputer's technique of using DeltaNet state as machine memory.
1 source
GitHub
OpenAI
00:00
OpenAI publishes excerpts of fair use briefs backing AI training
OpenAI shared excerpts from court briefs submitted by creators, academics, accessibility advocates, and legal experts arguing that fair use supports AI training and innovation. The move highlights the company's effort to build public and legal support for its position in ongoing copyright litigation over training data.
1 source
OpenAI
Meta
00:40
Meta quietly launches Style Sketch AI fashion app in Southeast Asia
Meta has launched Style Sketch, an AI app that helps users create outfit ideas from clothes they already own. The app appeared quietly in Indonesia, Singapore, and Thailand. It signals Meta's continued push into consumer AI products beyond its core social apps.
1 source
Techmeme
AI industry
20:40
Dartmouth Provost loses job after using AI despite warning against cheating
The provost of Dartmouth College was dismissed following the discovery of unauthorized artificial intelligence use in his work. This situation brings attention to the ethical challenges and accountability surrounding AI deployment, even among high-level administrators.
1 source
Ars Technica
AI industry
17:38
Starbucks plans AI-personalized drive-thru menus
Starbucks CEO Brian Niccol said during an earnings call that the company will explore recognizing customers at the drive-thru, presenting personalized digital menus, and making payment easier. The plan signals how major retail chains are bringing AI-driven personalization to in-person and drive-thru purchasing experiences.
1 source
The Verge
AI industry
13:00
Runware to mass-produce prefab AI data centers on assembly line
Startup Runware says it will build a factory that manufactures prefabricated AI data centers on an assembly line, aiming to produce them like cars. The plan could cut data center construction time and cost, though whether it can deliver at scale remains unproven.
1 source
TechCrunch
Nvidia
15:00
Nvidia details agentic AI helping the search for extraterrestrial signals
Nvidia described how agentic AI is being used to advance the search for extraterrestrial intelligence, in a Halloween-themed blog post. It notes SETI research increasingly uses AI agents built by humans to scan for signals, framing the work as machines listening for other machine intelligences.
1 source
Nvidia
Anthropic
14:22
Anthropic restricts AI rewrites of video game content
Reports say Anthropic is taking action to limit AI-generated rewrites of video game content. The move suggests the company is tightening its usage policies around creative works. It signals growing attention to how AI tools interact with games.
1 source
twitter.com
Policy & law
14:58
Pope calls for AI disarmament and ban on lethal autonomous weapons
Pope Leo urged United Nations member states to move toward banning lethal autonomous weapons. He asserted that AI systems function strictly as tools rather than as autonomous entities. This call underscores the importance of human guidance and responsibility in AI development.
1 source
The Verge
AI industry
10:10
Ravical raises $17M seed for AI platform serving accounting firms
Ravical, which builds AI software for accounting and tax firms to manage operations, raised a $17M seed round led by MMC. This brings its total funding to $25M. The startup aims to help firms turn individual expertise into shared knowledge as AI reshapes accounting.
1 source
Techmeme
Policy & law
11:00
Maine Senate race spotlights data center backlash
At the final Maine Senate debate, an audience member interrupted Republican Sen. Susan Collins to denounce AI data centers over climate and water impacts. Collins tied the outburst to her Democratic challenger. The incident shows AI infrastructure has become a live issue in midterm campaigns.
1 source
The Verge
AI industry
22:56
Privacy-focused AI assistant Ollie to shut down next week
Ollie, an AI assistant built around user privacy, is shutting down within a week, according to its founder. The closure comes as the AI assistant market grows increasingly crowded with competitors.
1 source
TechCrunch
AI industry
16:03
Theranos whistleblower Tyler Shultz warns AI boom invites deception
Theranos whistleblower Tyler Shultz said the AI boom is a great opportunity to deceive people, drawing parallels to the Theranos fraud. He warned that unprofitable companies raising capital at trillion-dollar valuations echo the culture that enabled Theranos. The cautionary tale highlights risks of hype-driven investing in AI.
1 source
TechCrunch
Nvidia
17:57
NVIDIA engineer builds Halloween talking garage door with Nemotron stack
An NVIDIA employee built MonsterHouse, an interactive Halloween projection where a grumpy pirate-ship-themed garage door talks with trick-or-treaters. The demo uses the Nemotron Voice Agent pipeline, Nemotron ASR, the Bolt E4B language model, Magpie TTS Zeroshot and Audio2Face-3D on a single workstation. It shows the voice agent stack running fully self-contained for interactive characters.
1 source
Hugging Face
AI industry
15:52
AI is deskilling site reliability engineers, report argues
A piece on the SigNoz newsletter argues that AI tools are not replacing site reliability engineers but instead deskilling them, as automation of on-call and debugging tasks erodes hands-on expertise. The article raises concerns about long-term skills loss in operations teams as AI assistance becomes routine.
1 source
newsletter.signoz.io
AI industry
15:00
Ghost hunters turn to AI chatbots in paranormal investigations
The Guardian reports that ghost hunting has shifted with AI chatbots becoming part of investigators' toolkits. Paranormal researchers Amy Waine and Jarrad Cutting, who build their own sensor-based apps, are now exploring how tools like ChatGPT are being used in the field, a trend that has concerned some enthusiasts.
1 source
The Guardian
Google
16:14
Google rolls out 42 fixes for Home app and devices
Google announced 42 updates to its Home app and compatible speakers and cameras. The changes address common user annoyances with the smart home ecosystem. It is a modest quality-of-life update rather than a new product launch.
1 source
ZDNET
AI industry
13:11
The Verge examines the risks of investing in speculative AI promises
The Verge published a commentary on the difficulty of investing in AI companies whose futures range from transformative to potentially catastrophic for humanity. The piece argues that stock bets on AI resemble gambling on outcomes that even experts cannot predict. It highlights the tension between miracle claims and existential risk warnings in the sector.
1 source
The Verge
Google
18:18
Android 17 adoption reported at just 1%
A report says Android 17 has reached only about 1% of Android devices, with Android 16 present on roughly a quarter of phones. The slow rollout highlights Android's persistent fragmentation problem, leaving most users without the latest security and feature updates.
1 source
ZDNET
AI industry
Developing · since Thu
Startups bet AI can make formal software verification practical
A new startup called &Truth has launched with $70 million in funding to automate formal verification, a mathematical method of proving software correctness. Rivals including Harmonic and Anthropic-backed efforts are also pushing the field. Advocates believe AI could extend verification beyond critical systems like aviation and cloud infrastructure.
1 new · 3 sources
Techmeme
The Guardian
Wired
Anthropic
Developing · since Thu
Anthropic renames subscription tiers and removes usage limits for Ultra users
Anthropic updated its premium subscription services by rebranding its higher tiers and increasing usage limits. The company is also warning stakeholders about geopolitical risks tied to its dependency on Taiwan for AI chip production.
1 new · 2 sources
Techmeme
The Verge
AI industry
Developing · since Thu
Dream raises $17M seed led by a16z for construction AI
Dream, founded by ex-Palantir executives Andy Alder and Arjun Mather, emerged from stealth with a $17 million seed round led by Andreessen Horowitz. The startup trains AI world models on US building permits and construction data to automate building design and engineering, aiming to challenge Autodesk.
1 new · 2 sources
Techmeme
TechCrunch
AI industry
Developing · since Wed
Kentucky administrator wins religious exemption from using AI at work
University of Kentucky administrator David Durman has received a religious exemption from using AI in the workplace. He is reportedly the first humanist granted such an exemption. The case may shape how AI-related conscience claims are handled.
1 new · 2 sources
TechCrunch
religionnews.com