THE ONGOING STORY

AI news, model releases and research

Follow AI model launches, research, agents and the companies building them. Catch up on a chosen period and compare the reporting behind each development.

AI illustration

Top in AI

Ranked by freshness, independent coverage and momentum, with a mix of publishers.

For you →Since your last visit →Saved stories →Interests and saves stay on this device.

ChatGPT flaw let a shared link read your private chats

Opening a stranger's ChatGPT share link was enough for an attacker to read your private conversations and act in your account, security firm Parameter reports. OpenAI fixed the flaw on October 26, and users don't need to install anything. Here is how to review and delete the links you have shared yourself.

via Notebookcheck

Sam Altman criticizes consumer PC memory usage as OpenAI pursues more data center chips

OpenAI Chief Executive Sam Altman sparked consumer backlash after arguing that personal computers underutilize scarce memory chips. Speaking on the Cheeky Pint podcast alongside Stripe co-founder John Collison, Altman said that from a global allocation perspective, it is "quite terrible" for limited DRAM supplies to sit inside individual PCs and run only 5% of the time, contending that cloud data centers maximize utilization far better for workloads like AI inference.

Summarized from 8 outlets
Read the full story

Arrowmax launches a camera-equipped smart screwdriver on Kickstarter that recognizes fasteners

Hong Kong-based toolmaker Arrowmax has launched a Kickstarter campaign for the Arrowspace S1, an electric screwdriver that relies on an integrated camera and artificial intelligence to identify fasteners and adjust torque. Surpassing its original target of just under $11,500, the project has collected more than $164,000. Kickstarter pledges start at around $79 or $80 for the standalone screwdriver, while a complete kit featuring a Smart Case and 52 bits made from S2 steel runs about $129, with deliveries scheduled to begin in April 2027.

Summarized from 2 outlets
Read the full story

Federal judge refuses to dismiss environmental lawsuit against Elon Musk's xAI

US District Judge Debra M. Brown has rejected a motion by xAI to dismiss a lawsuit brought by the NAACP over emissions at its Colossus data center outside Memphis. Filed in April with representation from the Southern Environmental Law Center, the legal action alleges that the company installed and ran at least 35 unpermitted combustion turbines and other pollution sources, violating the Clean Air Act.

Summarized from 2 outlets
Read the full story

Google introduces Gemini 4 Carbon frontier model with enhanced security tools

Google has announced Gemini 4 Carbon, a new flagship artificial intelligence model arriving just two weeks after the launch of Gemini 4 Argon. The system targets complex programming and security analysis, with pricing set at $2 per million input tokens, $10 per million output tokens, and a 95% discount for cached inputs. Carbon is not yet open to the general public, launching first to a limited group of cybersecurity defenders.

Summarized from 11 outlets
Read the full story

Mistral releases open weights for its trillion-parameter Large 4 model

French artificial intelligence lab Mistral AI has published the open weights for Mistral Large 4, a trillion-parameter multimodal model nicknamed "Le Chonk." The release on October 29, 2026, follows a three-week preview rollout that began on October 6. Outlets report varying figures for its active parameters, with SiliconANGLE citing 52 billion and Hugging Face noting 49 billion. According to Mistral, the model was trained from scratch in its European data centers over two months using 4,000 Nvidia Grace Blackwell GPUs.

Summarized from 5 outlets
Read the full story

American intelligence targets Singapore cloud executive in suspected operation directed at Nvidia CEO

United States intelligence agencies have placed Alice Huang, also known as Huang Le, under surveillance over suspicions that she was working as a Chinese intelligence asset and targeting Nvidia Chief Executive Jensen Huang. The New York Times reported that officials monitored the Singapore-based Megaspeed executive amid broader scrutiny over alleged diversions of advanced Nvidia chips to China.

Summarized from 7 outlets
Read the full story

Compromised Claude Code extension on Cursor marketplace exposed thousands of developers to an SSH backdoor

A tampered version of Anthropic's Claude Code extension slipped onto Cursor's marketplace on October 29, exposing up to 5,000 developers to malware designed to plant an SSH backdoor. Anyone running the AI code editor with automatic updates enabled could have received the package without manual intervention.

Summarized from 3 outlets
Read the full story

Open-source boot utility lifts AI processing limits on Nvidia GeForce cards

A community-developed open-source project named nv-uefi-unlocker allows owners of Nvidia GeForce RTX 20, 30, and 40-series graphics cards to bypass built-in Tensor Core restrictions on Windows and Linux. The utility boots from a FAT32 USB drive prior to operating system initialization, lifting throttles on tensor processing units without flashing video BIOS or modifying drivers. In benchmarks highlighted by Tom's Hardware, a laptop RTX 4060 achieved 65 percent higher FP16 math performance, while Wccftech notes the utility promises up to 75 percent higher AI compute performance across supported hardware.

Summarized from 2 outlets
Read the full story

Apple upgrades entry-level MacBook Pro and iMac with new M6 silicon

Apple has refreshed its entry-level 14-inch MacBook Pro and 24-inch iMac with its new 2nm M6 chip. The updated 14-inch MacBook Pro starts at $1,999, while the iMac begins at $1,499 for a base 10-core CPU and 10-core GPU model, rising to $1,699 for a 12-core edition. Pre-orders open on October 28, with retail availability and shipping scheduled to start on November 6.

Summarized from 10 outlets
Read the full story

AI briefings

· 2 source reports

Arrowmax launches a camera-equipped smart screwdriver on Kickstarter that recognizes fasteners

Hong Kong-based toolmaker Arrowmax has launched a Kickstarter campaign for the Arrowspace S1, an electric screwdriver that relies on an integrated camera and artificial intelligence to identify fasteners and adjust torque. Surpassing its original target of just under $11,500, the project has collected more than $164,000. Kickstarter pledges start at around $79 or $80 for the standalone screwdriver, while a complete kit featuring a Smart Case and 52 bits made from S2 steel runs about $129, with deliveries scheduled to begin in April 2027.

AI-generatedRead briefing and sources

· 2 source reports

Federal judge refuses to dismiss environmental lawsuit against Elon Musk's xAI

US District Judge Debra M. Brown has rejected a motion by xAI to dismiss a lawsuit brought by the NAACP over emissions at its Colossus data center outside Memphis. Filed in April with representation from the Southern Environmental Law Center, the legal action alleges that the company installed and ran at least 35 unpermitted combustion turbines and other pollution sources, violating the Clean Air Act.

AI-generatedRead briefing and sources

· 9 source reports

Sam Altman criticizes consumer PC memory usage as OpenAI pursues more data center chips

OpenAI Chief Executive Sam Altman sparked consumer backlash after arguing that personal computers underutilize scarce memory chips. Speaking on the Cheeky Pint podcast alongside Stripe co-founder John Collison, Altman said that from a global allocation perspective, it is "quite terrible" for limited DRAM supplies to sit inside individual PCs and run only 5% of the time, contending that cloud data centers maximize utilization far better for workloads like AI inference.

AI-generatedRead briefing and sources

· 2 source reports

Open-source boot utility lifts AI processing limits on Nvidia GeForce cards

A community-developed open-source project named nv-uefi-unlocker allows owners of Nvidia GeForce RTX 20, 30, and 40-series graphics cards to bypass built-in Tensor Core restrictions on Windows and Linux. The utility boots from a FAT32 USB drive prior to operating system initialization, lifting throttles on tensor processing units without flashing video BIOS or modifying drivers. In benchmarks highlighted by Tom's Hardware, a laptop RTX 4060 achieved 65 percent higher FP16 math performance, while Wccftech notes the utility promises up to 75 percent higher AI compute performance across supported hardware.

AI-generatedRead briefing and sources

· 2 source reports

Anthropic engineering manager drops traditional coding tests after finding Claude outperforms candidates

Felix Rieseberg, an engineering manager at Anthropic who has led work on Claude.ai and the Claude Code desktop app, has eliminated traditional coding interviews from his team's hiring loops. Rieseberg announced the move in a blog post, arguing that putting applicants through live technical tests no longer reveals anything meaningful about their capabilities because Claude now outperforms human candidates at those tasks.

AI-generatedRead briefing and sources

· 2 source reports

Justice Department indicts five people in nationwide student loan fraud investigation

The U.S. Department of Justice has indicted five individuals as part of Operation Ghostbusters, a nationwide crackdown targeting a widespread student loan fraud scheme. Investigators say bad actors used stolen identities and artificial intelligence to apply for federal student aid, creating ghost students who appeared to attend online college courses. Estimates suggest the activity has cost the U.S. government over $200 million.

AI-generatedRead briefing and sources

· 2 source reports

Decentralized network Darkbloom cautions Mac owners that falling token prices threaten compute earnings

Darkbloom, a distributed compute project run by Seattle-based Eigen Labs, is warning roughly 800 participants to expect sharp drops in revenue from renting out their idle Mac hardware. For example, Anthropic token prices fell 77% from $2.65 per million on May 1, 2026, to $0.62 on October 22, 2026, which would push monthly earnings down to roughly $36.77.

AI-generatedRead briefing and sources

· 2 source reports

Intel redirects chip fabrication capacity from consumer PCs to AI server processors

Intel is reallocating manufacturing capacity away from consumer PC processors to focus on server chips, responding to surging demand for artificial intelligence hardware. Chief executive Lip-Bu Tan described the crunch as one of the most significant supply shortages in history. The company's Data Center and AI division posted $7.3 billion in revenue for the third quarter of 2026, marking a 17% sequential increase and a nearly 80% jump compared to the previous year. Intel expects client PC demand to keep falling, while demand for data center processors is projected to stay strong through 2027 and 2028.

AI-generatedRead briefing and sources

· 2 source reports

Cisco tasks a 27-year-old engineer with an ambitious two-year legacy code overhaul powered by artificial intelligence

Cisco President and Chief Product Officer Jeetu Patel has granted 27-year-old employee Padmanabhan Krishnamurthy unlimited artificial intelligence tokens to lead a massive software overhaul. Krishnamurthy, the cofounder of translation startup EzDubs, joined Cisco through an acquisition and initially proposed refactoring 98% of the company's legacy code with just one other person. The project kicks off with a four-week demonstration period, aiming to complete the transformation within 24 months.

AI-generatedRead briefing and sources

· 4 source reports

Fujifilm commits 10 billion yen to expand magnetic tape manufacturing for artificial intelligence archives

Fujifilm is investing 10 billion yen (roughly $63 million to $64 million) to multiply its magnetic tape manufacturing capacity fivefold by fiscal 2028. The multi-year project at its Odawara facility within the Kanagawa Factory runs from fiscal 2026 through 2028 and represents the company's largest tape storage investment in approximately two decades. Output will concentrate on advanced cartridges capable of holding up to 100TB each, primarily latest-generation FUJIFILM LTO Ultrium10 media, measured against fiscal 2025 production baselines.

AI-generatedRead briefing and sources

· 2 source reports

Apple prepares to test localized Apple Intelligence in China ahead of a December release

Apple will begin testing Apple Intelligence in China on November 4 through its next iOS developer beta, ahead of an official public debut scheduled for December alongside the release of iOS 27.3. The launch follows a prolonged regulatory review process in the country, during which Chief Executive Tim Cook confirmed that the platform had secured regulatory approval.

AI-generatedRead briefing and sources

· 5 source reports

Mistral releases open weights for its trillion-parameter Large 4 model

French artificial intelligence lab Mistral AI has published the open weights for Mistral Large 4, a trillion-parameter multimodal model nicknamed "Le Chonk." The release on October 29, 2026, follows a three-week preview rollout that began on October 6. Outlets report varying figures for its active parameters, with SiliconANGLE citing 52 billion and Hugging Face noting 49 billion. According to Mistral, the model was trained from scratch in its European data centers over two months using 4,000 Nvidia Grace Blackwell GPUs.

AI-generatedRead briefing and sources

· 2 source reports

DoorDash CEO Tony Xu reveals consumers are barely using automated software to place delivery orders

DoorDash co-founder and chief executive Tony Xu revealed during a third-quarter earnings call on Thursday that third-party AI agents have generated only several hundred orders across the entire time they have been operational. Xu described the activity as a very low use case compared to the delivery platform's overall scale, which approached 1 billion total orders in the third quarter of 2026 alone.

AI-generatedRead briefing and sources

· 4 source reports

University of Kentucky administrator secures religious exemption to opt out of workplace AI

David Durman, a 45-year-old research administrator at the University of Kentucky, has been granted a workplace religious accommodation to avoid using artificial intelligence tools. Supported by the American Humanist Association, Durman based his request on Title VII of the Civil Rights Act of 1964, arguing that outsourcing his thinking to software violates his secular humanist convictions regarding human flourishing and agency. American Humanist Association staff attorney Alexandria Huiskens described the decision as the first of its kind for humanism.

AI-generatedRead briefing and sources

· 2 source reports

Accenture defense contractor linked personal Claude agent to secure Pentagon network

An Accenture contractor working for the Department of Defense granted a personal Claude AI agent access to a secure Pentagon environment to write and deploy code. The Claude Code agent remained connected to the Defense Department's Non-classified Internet Protocol Router Network (NIPRNet) for two months before officials detected the risk and patched the vulnerability.

AI-generatedRead briefing and sources

· 2 source reports

Finnish authorities open criminal investigation into Google over AI site forest clearings

Finnish police have launched a criminal investigation into Google subsidiary Tuike Finland Oy over suspected environmental damage and regulatory breaches at an artificial intelligence data center site in Muhos. Authorities initiated the probe following five complaints, backed by Finland's Licensing and Supervision Authority, after the company cleared and drained 330 hectares (816 acres) of forest before completing required environmental impact assessments. Construction work at both the Muhos location and a second project in Kajaani-Otanmäki halted on October 14.

AI-generatedRead briefing and sources

· 2 source reports

Artificial intelligence hardware boom strains circuit board supply through 2030

Booming demand for artificial intelligence hardware is severely constraining the global supply of printed circuit boards, pushing order books out as far as 2030, according to Nikkei Asia. Hardware manufacturers across the personal computing industry now face mounting challenges securing basic electronic components as high-paying AI infrastructure buyers command factory capacity.

AI-generatedRead briefing and sources

· 2 source reports

Elon Musk's data center firm moves to purchase neighborhood properties conditioned on waiving lawsuits

Elon Musk's data center business, SpaceXSI, is seeking to purchase 418 homes in the Colonial Hills neighborhood of Southaven, Mississippi. According to The Guardian, real estate agents have visited residents door to door offering buyout deals. However, sales are contingent on homeowners agreeing not to join an ongoing class-action lawsuit concerning noise pollution generated by the gas turbines powering nearby AI data centers.

AI-generatedRead briefing and sources

· 2 source reports

Meta tests Style Sketch to turn social media photos into digital wardrobes

Meta has rolled out an AI-powered wardrobe application called Style Sketch on Android and iOS across Indonesia, Singapore, and Thailand. First spotted by social media consultant Matt Navarra, the tool scans images from connected Facebook and Instagram accounts to build a digital closet, though users can also manually snap new pictures, record short videos, or upload camera roll photos.

AI-generatedRead briefing and sources

· 2 source reports

Dayton orders removal of Flock surveillance cameras reinstalled without city approval

Flock has reinstalled two surveillance cameras in Dayton, Minnesota, despite explicit city directions halting the program. Local officials stated on Facebook that the Dayton Police Department received a maintenance work order from Flock on September 17, 2026, prompting the city to warn the company the following day that the program was frozen and no work should proceed. Nevertheless, the cameras were mounted without local authorization, leading residents to discover and report the equipment.

AI-generatedRead briefing and sources

· 17 source reports

Apple upgrades entry-level MacBook Pro and iMac with new M6 silicon

Apple has refreshed its entry-level 14-inch MacBook Pro and 24-inch iMac with its new 2nm M6 chip. The updated 14-inch MacBook Pro starts at $1,999, while the iMac begins at $1,499 for a base 10-core CPU and 10-core GPU model, rising to $1,699 for a 12-core edition. Pre-orders open on October 28, with retail availability and shipping scheduled to start on November 6.

AI-generatedRead briefing and sources

· 7 source reports

American intelligence targets Singapore cloud executive in suspected operation directed at Nvidia CEO

United States intelligence agencies have placed Alice Huang, also known as Huang Le, under surveillance over suspicions that she was working as a Chinese intelligence asset and targeting Nvidia Chief Executive Jensen Huang. The New York Times reported that officials monitored the Singapore-based Megaspeed executive amid broader scrutiny over alleged diversions of advanced Nvidia chips to China.

AI-generatedRead briefing and sources

· 3 source reports

Strava links user activity history directly to Google Gemini for conversational workout analysis

Strava is expanding its artificial intelligence features by allowing paying subscribers to connect their accounts directly to Google Gemini. The rollout is currently limited to subscribers in the United States, with no confirmed schedule yet for availability in other countries. The feature is open to Strava subscribers regardless of which Gemini tier they use.

AI-generatedRead briefing and sources

· 4 source reports

Amazon instructs developers to stop inflating artificial intelligence usage

Amazon told its software development engineers this week to curb "tokenmaxxing," warning that running up high artificial intelligence usage does not equate to productivity. In an internal memo from Amazon distinguished engineer Terk Terkowitz published on October 27, engineering leadership cautioned against treating token consumption, AI adoption rates, code changes, and code reviews as primary yardsticks of success. While Amazon affirmed that artificial intelligence will remain central to software engineering, leadership stressed that teams should evaluate performance by customer impact and long-term results rather than raw input volume.

AI-generatedRead briefing and sources

· 11 source reports

Google introduces Gemini 4 Carbon frontier model with enhanced security tools

Google has announced Gemini 4 Carbon, a new flagship artificial intelligence model arriving just two weeks after the launch of Gemini 4 Argon. The system targets complex programming and security analysis, with pricing set at $2 per million input tokens, $10 per million output tokens, and a 95% discount for cached inputs. Carbon is not yet open to the general public, launching first to a limited group of cybersecurity defenders.

AI-generatedRead briefing and sources

· 2 source reports

Amazon shares slide after heavy spending on artificial intelligence triggers quarterly misses

Amazon missed analyst expectations for third-quarter revenue and earnings per share, sending its shares down more than 7% in late trading. Net sales rose 12% to $201.9 billion, falling just short of the $202.2 billion expected by analysts, while fully diluted earnings came in at $1.89 per share compared to estimates of $1.97. However, Amazon Web Services outperformed projections, expanding 39% year over year to reach $45.98 billion in sales and $17.8 billion in operating income.

AI-generatedRead briefing and sources

REFERENCE REVIEW · November 1, 2026

What changed in the model lineup

Compare models →

Gemini 4 Carbon

Announced October 28 for long-horizon coding and cybersecurity work. It leads the reviewed v5.0 index, but Google is initially limiting access to selected cyber defenders through Fairwind.

Provider details ↗

GPT-6.1 Terra

A lower-cost option for visual reasoning, computer use and writing. Tool calling uses the Responses API; evaluate screen-based tasks separately from the general Intelligence Index.

Provider details ↗

Kimi K3.1 Pro

The new K3.1 model supports vision and long-context agent work. Moonshot says full weights will be released by November 4; that release is still pending at this review. The Highspeed variant is a separate offering.

Provider details ↗

The benchmark reference now uses Artificial Analysis Intelligence Index v5.0. Its revised tests and weights make earlier index scores incompatible. See dated results and methodology →

What to look for

For a model announcement, the useful questions are what changed, who can use it, what access costs, and which performance claims have independent support. Research results, product announcements and broad availability describe different stages of progress.

Briefings are generated automatically from source excerpts with citations. The saved archive covers collected reports and may have gaps. How it works.