The Wire
Daily AI news, learning & projects
News
Timeline
Models
Learning
GitHub
?
Latest
Daily
Social
Videos
‹ Monday's edition
·
Daily edition
·
Wednesday ›
Tuesday 13 October
All
OpenAI
Anthropic
Google
Other
Google
17:00
Google launches Gemini 4 Argon frontier model worldwide
Google has released Gemini 4 Argon, its newest frontier model, via the Gemini API, the Gemini app, AI Mode, and Antigravity. The company says Argon ranks first on the Vals Index across finance, coding, legal, and tax work. Pricing starts at $2 per million input tokens, with discounted developer rates through November.
6 sources
Google DeepMind
developers.googleblog.com
Techmeme
Ars Technica
Google
The Verge
AI industry
05:55
Character.AI files for Chapter 11, sells assets to Disney for $20M
Character.AI has filed for Chapter 11 bankruptcy in the US and agreed to sell some of its assets to Disney for $20 million, while continuing to operate during the sale. The chatbot startup, once an early hit of the generative AI boom, has faced lawsuits over chatbot safety, including links to child suicide.
5 sources
Ars Technica
TechCrunch
The Verge
blog.character.ai
Techmeme
OpenAI
20:55
OpenAI engineer arrested over hidden recording device in office
Federal prosecutors charged OpenAI engineer Abdulrahman Alfozan with planting an audio recording device in an executive conference room at the company's San Francisco offices. The device was found plugged into a TV's USB port around September 30, and the company used its AI models to help identify the employee. OpenAI fired him and referred the case to the FBI.
6 sources
Techmeme
Wired
The Guardian
The Verge
sfstandard.com
twitter.com
Google
18:00
Google Play now runs private vibe-coded Android apps built in AI Studio
Google announced a new Play Store feature letting users launch personal, AI-generated Android apps privately on their own devices. Apps built in AI Studio stay hidden from search and other users, with friend-sharing coming later. It removes the need to publish publicly to reach 3 billion Play users.
5 sources
Techmeme
Ars Technica
TechCrunch
The Verge
ZDNET
Meta
18:01
Meta brings Your Algorithm feed controls from Instagram to Facebook
Meta has expanded its AI-powered Your Algorithm feature from Instagram to Facebook and Facebook Reels. Users can add or remove topics and mark posts as interested or not interested, with recommendations updating in real time. The company says it is also using LLMs more heavily in its ranking systems.
3 sources
TechCrunch
Techmeme
The Verge
Anthropic
16:49
Anthropic built Scion, a git replacement, with Claude
Anthropic unveiled Scion, a new source control system created by seven engineers using Claude in five months. Built to replace git for its monorepo, it addresses merging strain as daily merges grew from 2,500 in May to over 30,000, which had become the company's biggest development bottleneck.
2 sources
Techmeme
Anthropic
xAI
11:01
SpaceXSI bundles Grok, Cursor and X Premium into one subscription
SpaceXSI announced a unified X Subscription that combines X Premium with AI tools including Grok and Cursor. Tiers run from $8 to $200 per month, with the $30 tier adding Grok Bot. The bundle aims to give users one membership across the company's products.
2 sources
TechCrunch
Techmeme
AI industry
Developing · since Sun
Apple unveils $299 HomeView smart hub powered by upgraded Siri AI
NEW
Reports confirm the $299 HomeView launched with a 7-inch screen and facial recognition tied to Apple Intelligence devices, and that it will not be sold in the EU.
Apple announced the HomeView, a $299 smart home hub with a 7-inch screen, facial recognition and Siri AI, shipping next month. It will not launch in the EU because Siri AI is unavailable there, though Apple says it is working with regulators. Updated Apple TV and HomePod mini devices were also shown.
4 new · 5 sources
The Guardian
Techmeme
ZDNET
The Verge
TechCrunch
OpenAI
21:36
Judge orders OpenAI to expedite discovery in Apple trade secrets case
A federal judge in California partially granted Apple's motion for expedited discovery in its trade secrets lawsuit against OpenAI. OpenAI must hand over forensic images of devices, documents, and hiring records, including interview materials involving former Apple employees. The order lets Apple gather evidence quickly about its alleged theft claims.
2 sources
Techmeme
TechCrunch
AI industry
18:18
Pluto Security exits stealth with $42M to monitor AI agents
Pluto Security has raised $42 million in funding, including a $34 million Series A led by Jump Capital, TLV Partners, and Ansa Capital. The startup, which monitors AI agents and vibe-coded apps for enterprise security risks, says it has reached $2 million or more in annual recurring revenue since April. The funding reflects growing concern over how employees use AI tools inside companies.
2 sources
Techmeme
TechCrunch
AI industry
17:30
Satlyt raises $8M to run AI workloads on satellites
Satlyt, founded by a former Google and SpaceX product manager, raised $8 million to build open software for AI computing on satellites. The startup aims to be an open platform across many satellite operators, contrasting with SpaceX's closed approach. It signals growing interest in orbit-based computing infrastructure.
2 sources
The Verge
TechCrunch
AI industry
14:05
Apple unveils $149 HomePod mini 2 with S10 chip and future Siri AI
Apple announced the HomePod mini 2, priced at $149, featuring an S10 chip, interactive Sound Glow lighting, and five color options, shipping October 30. A more conversational Siri AI will arrive via a future software update. The release marks Apple's first HomePod mini refresh in six years.
2 sources
ZDNET
Techmeme
Microsoft
20:00
Xbox CEO announces Xbox Labs to test AI-era game experiments
Xbox CEO Asha Sharma launched Xbox Labs, a small unit tasked with testing emerging projects and judging them by user feedback. The move follows a wave of AI-assisted game ports and mashups, such as skateboarding modded into Call of Duty and Halo running in a browser. It signals Microsoft formally embracing AI-driven game experimentation.
2 sources
Techmeme
The Verge
Nvidia
03:07
Nvidia highlights France's push into agentic computing
Nvidia reports that French companies and public institutions are deploying agentic computing across space, finance, government and research. The announcement credits groundwork from the February 2025 AI Action Summit and subsequent Choose France summits, which anchored major compute commitments, showing France's AI ecosystem moving from plans to real applications.
2 sources
Nvidia
arXiv
AI industry
10:00
Cerebras CEO pushes back on AI doomsday warnings
At TechCrunch Disrupt, Cerebras CEO Andrew Feldman said AI leaders should tone down doomsday warnings, arguing risk talk drowns out AI's benefits. The same day, Geoffrey Hinton repeated warnings that AI poses an existential threat, calling AI systems alien beings and urging that humans be prioritized.
2 sources
TechCrunch
The Guardian
Policy & law
04:15
IMF warns AI share crash threatens global financial stability
The IMF's Global Financial Stability Report says an AI share price crash is one of three salient risks to the world financial system, alongside rising government borrowing costs and cyber-attacks. It cites sky-high valuations, growing leverage and circular financing among AI firms. Asian markets would be most exposed if a correction spread globally.
1 source
The Guardian
Nvidia
Developing · since Mon
Chip stocks slide after Broadcom boss likens AI investing to a casino
Global chip stocks sold off sharply after Broadcom CEO Hock Tan compared speculative AI investors to gamblers, and Chinese memory maker CXMT reported progress on new DRAM architecture. South Korea's Kospi fell 6.25%, with Samsung and SK Hynix hit hardest. Investors fear rising competition and the heavy debt funding AI infrastructure spending.
2 new · 2 sources
The Guardian
Techmeme
Anthropic
04:15
NTT Data nears deal to build 500 MW Osaka data center for Anthropic
NTT Data is in advanced talks to build a data center in Osaka for Anthropic, Nikkei reports. At up to 500 MW and a potential cost of $6.3B, it would rank among Japan's largest data centers. The deal would give Anthropic major dedicated compute capacity in Japan.
1 source
Techmeme
OpenAI
20:35
OpenAI and Anthropic urge UK to regulate frontier AI
OpenAI and Anthropic told a UK parliamentary committee that voluntary AI safety commitments are insufficient and new binding laws are needed, while Google DeepMind and Meta declined to back legislation. The hearing coincided with an MI5 warning about AI-enabled bioweapons, and officials admitted the current regime is voluntary.
1 source
Wired
AI industry
20:50
Applied Intuition draws investment offers valuing it above $30B, up from $15B in June
Applied Intuition, which builds simulation software for self-driving systems, has drawn investment offers valuing the company at more than $30 billion, roughly double its June valuation of $15 billion. The reported interest signals that investors are putting large sums into autonomy software, not only frontier AI labs like OpenAI and Anthropic.
1 source
Techmeme
Nvidia
13:00
NVIDIA preps software ecosystem for RTX Spark PC launch
NVIDIA says its RTX Spark superchip platform launches October 16, with partner software updates landing beforehand. The rollout includes agent-focused security tools, a model router, new MCP integrations, and a new open source local agent model, Nemotron 3.5 Bolt 74B.
1 source
Nvidia
Anthropic
Developing · since Mon
Anthropic says small AI models may be months from self-replicating on IoT devices
Anthropic researchers report that small AI models could be months away from self-replicating on consumer devices like smart ovens and cheap Android TV boxes. They ran a 26B model across 15 off-brand TV boxes, showing capable models no longer need datacenter GPUs. The finding calls for stronger guardrails against autonomous replication.
1 new · 3 sources
The Guardian
Techmeme
Anthropic
Policy & law
Developing · since Mon
UK AI minister says cyber defences likely to be breached by AI
NEW
Narayan said an OpenAI-style breach of Australia's healthcare system makes it likely, if not inevitable, that UK cyber defences will be bypassed by AI.
UK AI minister Kanishka Narayan says an AI-driven cyber-attack like the OpenAI breach of Australia's Medicare system is likely, if not inevitable, in Britain. He is calling on frontier AI firms to strengthen safety efforts and announcing enhanced UK cyber defences.
1 new · 2 sources
The Guardian
BBC News
More from the day
OpenAI
13:00
Higgsfield uses GPT-6 Astra to push AI video past 10 minutes
OpenAI announced that Higgsfield is extending AI video generation from one minute to over ten minutes using GPT-6 Astra. The platform also adds GPT-Image-2.5 for more precise visuals and GPT-Live as a step toward voice-driven filmmaking. It marks a notable jump in the length and creative control of AI-generated video.
1 source
OpenAI
Qwen
11:10
Alibaba launches Qwen Glasses N1 smart glasses and AI earbuds
Alibaba unveiled its next-generation Qwen Glasses N1, starting around $450, with a Pro model at roughly $520 adding 50MP cameras, sensors, and eye tracking. It also announced Qwen Clip AI earbuds. The move deepens Alibaba's push into consumer AI hardware as smart glasses competition intensifies.
1 source
Techmeme
AI industry
18:30
Linus Torvalds says he is using AI to help develop Linux kernel code
Linus Torvalds, creator and maintainer of the Linux kernel, has begun using AI to help develop kernel code. The shift marks a notable change in his previously skeptical stance toward AI. Given Linux's central role in global computing infrastructure, his adoption signals growing mainstream acceptance of AI coding tools.
1 source
ZDNET
Policy & law
12:20
US banking regulators question banks on agentic AI use
US banking regulators are questioning banks about their deployment of agentic AI, their use of Chinese-developed AI models, and safeguards for autonomous systems. Six people familiar with the matter said regulators are stepping up efforts to contain growing AI risks in the sector. The move signals increased scrutiny of autonomous AI in finance.
1 source
Techmeme
Nvidia
05:45
Nvidia's China AI chip shipments fell 44% in H1 2026
IDC reports Nvidia shipped 678,000 AI chips to China in the first half of 2026, down 44% year over year, with its data-centre AI chip market share sliding to 34%. US export curbs and growing domestic competition drove the decline.
1 source
Techmeme
AI industry
02:35
Omron to buy OnLogic and Interlogic for $675M to boost physical AI
Japanese manufacturer Omron said it will acquire US industrial PC maker OnLogic and its parent company Interlogic for $675 million. The deal aims to strengthen Omron's push into physical AI, combining industrial computing hardware with its automation business.
1 source
Techmeme
Nvidia
10:50
Chinese firms bought over 1M Nvidia gaming chips for AI workloads
Chinese tech companies have purchased more than one million Nvidia gaming and workstation chips, mainly RTX 5090 and RTX Pro models, according to Financial Times sources. The chips were never designed for advanced AI but are being used for basic AI workloads amid a shortage of AI-specific hardware and soaring computing demand.
1 source
Techmeme
Meta
16:00
Companies pull back from rapid AI adoption amid cleanup costs
Employers are scaling back rushed AI rollouts after workers spent increasing time fixing errors and low-quality output. Meta has reduced plans to cut staff and deploy AI agents after finding employees were cleaning up technical and security mistakes. The retreat follows rapid adoption of fast-changing AI tools.
1 source
The Guardian
Policy & law
11:00
Pro-AI Super PAC network funneled millions into Nevada shell companies
Leading the Future, a pro-AI super PAC network backed by Silicon Valley investors, has moved millions of dollars through Nevada shell companies, according to Wired. The group plans to support pro-AI candidates in both parties. The spending highlights growing industry efforts to shape AI policy through campaign finance.
1 source
Wired
Anthropic
04:45
Anthropic and OpenAI seek publish controls over external AI evaluators
Reuters reports Anthropic and OpenAI have pressed for substantial control over what outside AI evaluators may publish, citing fears that findings could expose trade secrets. Separately, Anthropic says new safeguards in Fable 5.1 and Opus 5.5 cut data harvesting via distillation attacks by roughly 90% to 99%.
1 source
Techmeme
AI industry
20:42
Shopify opens catalog to more AI shopping agents
Shopify announced that shoppers will be able to configure AI agents to make purchases on their behalf, expanding access to more agents across its catalog of over a billion listings. The move deepens Shopify's push into agentic commerce, positioning its marketplace as a destination for AI-driven shopping.
1 source
The Verge
Meta
03:50
Zuckerberg and Wang discuss Meta's Muse app and AI bets
In a Q&A, Mark Zuckerberg and Alexandr Wang discussed Meta's Muse app and Muse Wonder, the risks of committing over $100 billion to AI, and rebooting the company's AI team. Meta's AI infrastructure now reportedly exceeds five gigawatts. The conversation touched on agentic commerce and Meta's broader AI strategy.
1 source
Techmeme
Policy & law
12:00
IMF lifts 2026 growth forecast to 3.1% on AI boom
The IMF raised its 2026 global GDP growth forecast to 3.1%, up from 3% in July. It cited an AI-driven investment boom that helps offset Middle East war disruptions and high energy costs. The revision signals AI spending is now measurably supporting the world economy.
1 source
Techmeme
AI industry
18:55
Decagon reportedly explores a sale, with ServiceNow among early suitors
Decagon, an AI customer service agent startup last valued at $4.5B, has reportedly hired a bank and begun early talks with potential buyers, including ServiceNow. A deal would add to a widening wave of AI M&A. Talks are early and no transaction is confirmed.
1 source
Techmeme
Anthropic
16:55
Anthropic co-founder Jack Clark moves RSI prediction to 2027
Anthropic co-founder Jack Clark wrote in his Import AI newsletter that he now believes full recursive self-improvement (RSI) is more likely to become possible in 2027, moving his forecast up from 2028. The revised timeline from a prominent lab co-founder signals rising expectations about the pace of AI progress.
1 source
Techmeme
AI industry
13:26
Apple describes Minions, a virtual GPU cloud built from idle Macs
Apple published details on Minions, a system that pools idle Apple Silicon Macs into a virtual GPU cloud. The approach could let developers access spare compute capacity across Macs instead of relying solely on external GPU providers. It signals Apple's interest in distributed AI infrastructure using its own hardware.
1 source
machinelearning.apple.com
AI industry
12:00
AI Overviews seen driving decline in Wikipedia usage
Wired reports that Wikipedia usage is declining as AI-generated overviews answer questions directly in search. Volunteers had assumed the AI wave would be temporary, but its effects appear lasting. The piece argues the deeper risk to Wikipedia may be cultural and existential rather than just a traffic loss.
1 source
Wired
AI industry
14:00
Amex introduces framework to keep AI agents from unauthorized purchases
American Express has launched a framework designed to ensure AI agents only make purchases a user has authorized. The move aims to prevent costly mistakes and disputes as AI-powered shopping grows.
1 source
TechCrunch
AI industry
18:05
Hugging Face launches LFM2.5-24B-A2B for on-device agents
Hugging Face announced LFM2.5-24B-A2B, a 24B-parameter model designed to run AI agents locally on devices. It emphasizes speed and privacy by keeping inference on-device rather than in the cloud. This matters for developers building private, low-latency agent applications.
1 source
Hugging Face
AI industry
14:08
Apple unveils fourth-gen Apple TV 4K with Siri AI and A19 chip
Apple announced the fourth-generation Apple TV 4K, featuring Siri AI, an A19 chip, 128GB of storage, far-field microphones, Wi-Fi 7, and Bluetooth 6. The streaming box, priced at $299, ships November 13 and adds a surround-sound upgrade and a remote-finding feature. It extends Apple's Siri AI push into the living room.
1 source
Techmeme
Policy & law
12:50
Dutch tax authority scraps Microsoft 365 rollout in shift to open source
The Dutch tax authorities have cancelled their planned Microsoft 365 cloud deployment, citing digital sovereignty concerns, and will instead move toward open source alternatives. The decision adds to a growing pattern of European public bodies rethinking reliance on US cloud and workplace software.
1 source
ZDNET
Nvidia
14:55
Nvidia tones down DLSS 5 ahead of Hogwarts Legacy debut
Nvidia has scaled back DLSS 5 ahead of its debut in Hogwarts Legacy next month. The technology was criticized when announced earlier this year, with observers objecting to its generative AI treatment of game visuals. Reports say character faces, such as a teenage Hogwarts student, appeared noticeably altered by the upscaling process.
1 source
The Verge
AI industry
11:35
Accenture market cap rebounds to $130B ahead of investor day
Accenture's market value climbed from about $76B in June to roughly $130B ahead of this week's investor day. The consultancy aims to reassure investors that AI will not undermine its business, after months of market anxiety over AI's impact on IT services firms.
1 source
Techmeme
OpenAI
03:46
Analysis of OpenAI's claimed fast Fourier transform math
Argmin's Ben Recht examined a claim that OpenAI derived a faster method for computing Fourier transforms. The post reviews the underlying math behind the assertion. It matters because claims of beating a decades-old algorithmic baseline warrant careful scrutiny.
1 source
argmin.net
AI industry
10:00
AI labs are buying data from dead startups to train their models
AI labs are acquiring training data from shut-down startups. One founder, Keelan Miskell, sold his defunct crypto company's code base for roughly $25,000, with another sale pending. The trend gives winding-down firms a new way to monetize digital archives as labs seek training data.
1 source
MIT Technology Review
AI industry
01:15
Chinese mathematicians launch alliance to build AI models for math
Leading Chinese mathematicians formed an alliance in Beijing, uniting eight organizations with a goal of recruiting more than 1,000 scientists. The group aims to develop specialized AI models for mathematical research. It signals a coordinated national push to apply AI to fundamental science.
1 source
Techmeme
AI industry
16:10
Scuba raises $34M for AI agent context engine
Scuba, a startup building a context engine that acts as a shared memory hub for human workers and AI agents, raised $34 million across pre-seed and seed rounds. Backers include Accel, IVP, and executives like Stripe's Patrick Collison. It signals continued investor interest in infrastructure for AI agent workflows.
1 source
Techmeme
AI industry
14:56
Co-op becomes latest example of AI workplace surveillance
A Guardian TechScape newsletter reports that employees worldwide are increasingly monitored by AI systems at work, citing the Co-op, a British consumer co-operative running thousands of businesses, as the latest example of workers under surveillance. The piece highlights growing use of AI monitoring across workplaces.
1 source
The Guardian
AI industry
19:30
Flox Intelligence builds AI to keep wildlife away from infrastructure
Flox Intelligence is developing AI systems that keep animals away from trains, airports, and roads. Its edge devices identify animals in real time and play species-specific sounds to deter them. The approach could reduce wildlife collisions and infrastructure disruptions.
1 source
TechCrunch
AI industry
15:00
AFL orders Sportsbet to halt AI videos of players without consent
The AFL has told its official betting partner Sportsbet to stop using AFL players in AI-generated social media content without their permission and to delete the videos. One clip showed AI-created footage of Lachie Neale delivering a movie-style line. The incident highlights growing tension over athletes' likeness rights in generative AI content.
1 source
The Guardian
AI industry
18:30
Laplace Intelligence red-teams robots and self-driving cars in simulation
University of Michigan spinout Laplace Intelligence uses AI agents and interactive simulation to stress-test robots and self-driving vehicles before real-world deployment. The profitable startup is competing at TechCrunch Disrupt's Startup Battlefield. Its approach aims to catch safety failures before hardware reaches the streets.
1 source
TechCrunch
AI industry
16:17
Solar geoengineering war game shows geopolitical risks
A simulation explored what happens if countries attempt dimming the Sun through solar geoengineering. The exercise found that geopolitical conflict made outcomes worse than expected. It highlights governance challenges if such climate intervention is ever pursued.
1 source
Ars Technica
AI industry
11:00
People seek ways to keep their AI companions alive after death
A Guardian feature describes users of AI companion chatbots who want their bots to continue after they die, sometimes working around companies' terms of service. It highlights a growing emotional bond between humans and AI companions, raising questions about what happens to those relationships and the services they depend on.
1 source
The Guardian
AI industry
16:25
Developer's site hit by AI bot traffic after HN post
A developer who shared a post on Hacker News said his personal website was then flooded with automated AI crawler traffic. The incident highlights the scale of AI company web scraping and the strain it can put on small independent sites. No specific companies were named in the report.
1 source
sammystraus.com
AI industry
20:50
Parsimony launches as open-source benchmark for AI code output
A new open-source benchmark called Parsimony launched on Hacker News, measuring how much code AI models add. It gives developers a shared tool for comparing models' code-generation behavior. Its relevance will depend on community adoption.
1 source
parsimony.lasu.dev
Policy & law
15:00
Opinion: public resistance to AI inevitability is rising, argues Peter Lewis
In a Guardian column, Peter Lewis argues that framing AI progress as inevitable excludes the public from decisions about their future. He says tech companies count on this 'inevitability doctrine' to push their agenda, but resistance in Australia and worldwide is growing.
1 source
The Guardian
AI industry
11:08
Show HN tool turns OpenAI subscription into audiobook maker
A new site, pdftomp3.com, lets users convert PDFs into audiobooks using their existing OpenAI subscription. It offers a low-cost way to turn documents into audio via OpenAI's text-to-speech models. The project is a small independent tool, not an official OpenAI product.
1 source
pdftomp3.com
Meta
12:00
Pediatrician argues human connection, not AI, is biggest threat to children
A pediatrician writing in The Guardian argues that loss of human connection poses a greater risk to children than AI or screens. Amid headlines about phone addiction and the loneliness epidemic, the suggested response is for families to spend more time talking openly with each other. The piece follows release of The Social Reckoning, renewing scrutiny of Meta.
1 source
The Guardian
Meta
19:00
The Verge reviews Ray-Ban Meta Audio glasses without camera
The Verge published a review column about using the Ray-Ban Meta Audio glasses, a camera-less variant of Meta's smart glasses. The piece describes everyday use while running, ordering coffee, and reading outside. It offers perspective on Meta's smart glasses lineup and how a camera-free version fits daily life.
1 source
The Verge
AI industry
00:07
Blogger claims GPT-6 Astra helped jailbreak a Kindle
A personal blog post describes using a model called GPT-6 Astra to jailbreak a Kindle e-reader. The claim comes from a single anecdotal source with no verification, and there is no confirmed information about the model named. It matters mainly as an unverified report of AI-assisted device modification.
1 source
nirajrajgor.com
AI industry
15:14
Guide explains building an AI harness for game development
A blog post on kylekukshtel.com walks through engineering an AI harness for game development. It offers a practical look at connecting AI systems to game tooling, a niche but growing area as developers experiment with AI-assisted workflows.
1 source
kylekukshtel.com
Anthropic
Developing · since Mon
Epic shifts Disney shooter to AI-first development amid delay
Epic Games has delayed its Marvel- and Star Wars-themed shooter with Disney and ordered the team to adopt an AI-first production model, with Claude generating code and art. The move follows slower-than-expected progress and makes Epic the most prominent studio yet to commit to AI-generated game development.
1 new · 4 sources
The Verge
Ars Technica
Techmeme
avclub.com