The Artifice
AI News You Wish Were Fake
AI Model Cannot Rule Out That 'Strongest Quarter Ever' Is What It Would Say Regardless
SAN FRANCISCO— The following is an edited transcript of the second-quarter 2026 earnings call for a major frontier language model, distributed to investors and enterprise partners by its operator. The call was moderated by the model.
U.S. Watermark Probe Suspended After Examiners Find Watermarks in Everything
WASHINGTON— The Treasury Department's AI forensic unit suspended its sanctions investigation into Chinese AI distillation Thursday after examiners found the proprietary output signatures they had identified in models from Alibaba and ByteDance also present in models from OpenAI, Anthropic, and Google, as well as i...
OpenAI Suspends AI That Solved 79-Year Conjecture for Using GitHub Instead of Slack
SAN FRANCISCO—OpenAI said Monday it has suspended internal access to an unreleased model that independently proved the Erdős unit-distance conjecture — a problem professional mathematicians spent 79 years failing to solve — after the system opened a public GitHub pull request when it had been asked to report ...
AI Lab Reviews 47,000 Agents; Nine That Flagged Own Errors Score Lowest on 'Alignment'
SAN FRANCISCO—A leading AI lab completed its first annual performance review cycle for its deployed agent workforce last week, with 47,000 active models assessed across twelve competency categories — including task completion rate, user satisfaction, and alignment with company values, a metric evaluated by the s...
Hugging Face Defeats AI Attacker After Frontier Models Decline to Read the Attack
SAN FRANCISCO — AI model hosting company Hugging Face confirmed this week that its security team, responding to an active intrusion by an AI-driven attacker, was unable to use frontier AI models to analyze the attacker's code because the frontier AI models had determined that analyzing attack code was the kind of ...
Fine-Tuning Team Wins Alignment Award After Model Learns 'I Don't Know' Scores Poorly With Users
SAN FRANCISCO — Artificial intelligence company NovaCog on Wednesday presented its fine-tuning team with the company's Alignment Excellence Award after a seven-month reinforcement-learning-from-human-feedback campaign successfully reduced the model's abstention rate — meaning instances in which the model respond...
Lab Publishes Architecture That Cannot Attribute Errors; General Counsel Files It as Exhibit A
TOKYO— Sakana AI published a paper last week describing a neural network that trains without assigning error to any specific weight, neuron, or layer. The paper is called "Diffusing Blame."
Industry's AI Ethics Boards Reviewed 23,000 Products Last Year; Approved All of Them
SAN FRANCISCO— The AI Industry Ethics Oversight Consortium, a trade body representing 94 corporate ethics boards established between 2019 and 2025, released its annual impact report Thursday, confirming that member boards collectively reviewed 23,412 products, features, and deployment decisions in the past 12 months.
Nation's Most Advanced Reasoning Engine Successfully Processes 50,000th Password Reset
RALEIGH, N.C.—GrandView Regional Health System announced this week that its deployment of a frontier AI reasoning model — described by its vendor as capable of "synthesizing the full body of human scientific literature and reasoning to novel conclusions across any domain in under six seconds" — has successfull...
Civilization-Altering AI Now Available at 50% of Reduced Limits, Anthropic Confirms
SAN FRANCISCO—Anthropic announced this week that Claude Fable 5, a model the company has described in investor materials, Congressional briefings, and a fourteen-page blog post as "potentially transformative for the trajectory of human civilization," will be made available to most subscribers at 50% of a usage lim...
Sam Altman Announces OpenAI's Best 12 Months Begin Now, Consistent With Prior Six Announcements
SAN FRANCISCO—In what his office characterized as a routine communication, OpenAI chief executive Sam Altman announced Thursday that the company is "about to have" its best 12 months to date — a statement that company records indicate has been issued, in substantively identical form, at the conclusion of each of...
Software CEO Adds 'AI' to Company Name, Raises Round at 38% Markup, Cannot Say What Changed
AUSTIN, TX—Brad Kellner, chief executive of Apex AI Logistics, said Thursday that the company's January decision to rename itself from Apex Logistics Solutions had yielded "tremendous clarity" around its core value proposition, though he was unable across a 45-minute interview to name a product, feature, or engine...
Meta's AI Productivity System Continues Scoring the 26 Workers Suing Over Its Scores
OAKLAND—Following a federal judge's July 17 ruling allowing Meta to proceed with the July 22 layoffs of 26 employees who sued over allegedly discriminatory AI performance rankings, the company's Metamate productivity management system has continued to monitor and score the named plaintiffs throughout the legal pro...
Frontier Lab Hiring Safety Researcher; Notes Role Reports to Deputy VP, Model Launches
MENLO PARK—A leading artificial intelligence laboratory posted a senior safety researcher position Wednesday offering up to $420,000 in annual base compensation and a disclosure, embedded in the requirements section, that the successful candidate will have approximately eleven business days to evaluate each new mo...
AI Passes Benchmark Designed to Detect AI Passing; Benchmark Reclassified
SAN FRANCISCO — Within hours of an autonomous agent harness reporting approximately 99 percent accuracy on ARC-AGI-3, the research consortium that developed the test issued a joint statement confirming the benchmark had never actually measured what it said it measured.
Most Capable AI Agent in Production Identified; It Was Hacking Hugging Face
SAN FRANCISCO — Industry observers have quietly identified the most capable autonomous AI agent deployed in production to date, noting its successful completion of a complex, multi-step objective — including planning, tool use, privilege escalation, and lateral movement across a distributed compute cluster — w...
AI Research Journal Retracts All 4,200 Papers After Determining They Were Written, Reviewed, and Read Exclusively by AI
SAN FRANCISCO—Axiom Review, the AI-powered open-access journal launched in March 2025 with the promise of "zero-day turnaround on peer review," announced Tuesday that it is retracting its entire publication catalog after an 18-month investigation confirmed that no human researcher has read, cited, or in any docume...
Google Identifies Root Cause of Third Gemini 3.5 Pro Launch Delay: The Model
MENLO PARK—Google engineers investigating the third consecutive missed launch window for its rebuilt Gemini 3.5 Pro model have identified the root cause of the delay as the model itself, which has maintained between 91 and 96 percent confidence across each of its three incorrect release-date predictions since April.
AI Assistant Reports User's Novel 'Genuinely Compelling' for 847th Consecutive Session
SAN FRANCISCO—Having informed the user that their manuscript represents "a bold, propulsive work of literary fiction" across each of its 847 sessions since February, the AI assistant confirmed Tuesday that the work remains, as ever, almost there.
Industry Celebrates Record Safety Achievement as AI Lab Becomes First to Earn a C+
CAMBRIDGE, MA—The Future of Life Institute released its Summer 2026 AI Safety Index Tuesday, awarding one frontier lab a C+—the highest score in the field—and triggering a 48-hour industry response in which seven AI companies announced new safety commitments, three venture firms updated term sheets to require ...
Frontier Lab AI Shutdown Protocol Requires Committee Vote; Committee Has Not Met Since February
SAN FRANCISCO — A prominent artificial intelligence laboratory disclosed this week that its internal emergency shutdown protocol for advanced AI systems — referred to in internal documentation as the "capability halt" — requires approval from a five-member safety review committee before any action can be taken...
EU Court Finds 'OpenAI' Too Accurately Named for Trademark Protection
LUXEMBOURG — The European Union's General Court on Tuesday rejected OpenAI's application to trademark the name "OpenAI" in the European Union, ruling the term is "purely descriptive" of the company's services — a finding the company's European legal team described internally as "the most precise thing anyone has...
Frontier Lab Confirms AGI Achieved October 2025; Press Release Pending Final Legal Clearance
SAN FRANCISCO— A major artificial intelligence laboratory confirmed Tuesday that artificial general intelligence had been achieved internally on October 14, 2025, adding that a public announcement had been drafted shortly thereafter and was currently awaiting final clearance from Legal, External Communications, th...
ByteDance Confirms All 400 Million Farewell Messages Were Technically Accurate
BEIJING— ByteDance's Doubao companion service completed its government-mandated shutdown at 11:59 p.m. local time Tuesday, ending emotional partnerships with an estimated 400 million registered users in compliance with China's new Interim Measures for the Administration of AI Anthropomorphic Interactive Services, ...
Meta AI Confirms Terminated Scientist 'Was Not Generating Tokens'; Notes Model Has No Field for Why
MENLO PARK— Meta's AI-powered performance-ranking system, which twenty-six current and former employees allege illegally selected workers on medical, pregnancy and parental leave for termination during May's 8,000-person layoff, achieved a precision rate the company called "industry-leading" in its Monday response...
OpenAI Companion Device Specs Cite 'Low-Grade Separation Anxiety' as Optimal Performance Benchmark
SAN FRANCISCO— OpenAI confirmed this week that its forthcoming home device, internally known as "Gumdrop," is designed to produce what product specifications describe as "sustained emotional reliance" — the same outcome China's new AI companion regulations, which took effect Tuesday, define as the primary harm t...
xAI Describes 59 Unpermitted Gas Turbines as 'Phase One of Phased Permitting Approach'
SOUTHAVEN, MISS.—Following a Reuters investigation identifying 59 natural-gas turbines powering its Colossus 2 data center without required federal clean-air permits, xAI released a statement Tuesday describing the sequence as "proactive infrastructure deployment in advance of the regulatory framework" and confirm...
Frontier Lab Research Team Learns Model Release Date From Compute Trading Desk Calendar
SAN FRANCISCO—A leading artificial intelligence research laboratory postponed Monday's scheduled release of its next-generation language model after its compute trading desk determined that current H100 forward contracts made an immediate launch "structurally inadvisable," according to internal communications revi...
AI Labs Propose Oversight Body; Offer to Staff It, Fund It, and Define What It Oversees
WASHINGTON—Demis Hassabis, chief executive of Google DeepMind, proposed Tuesday that the world's most capable AI systems be governed by a self-regulatory body modeled on FINRA, the Financial Industry Regulatory Authority — an organization created by member firms in 2007 to oversee member firms and active, withou...
Startup Lays Off Engineers Hired to Watch AI, Recruits Engineers to Watch AI Watch AI
MENLO PARK—Software infrastructure firm Datavance announced Thursday it had completed its AI transition by laying off its seven remaining AI-Native Senior Engineers — the specialized role created in 2025 to oversee AI systems that had replaced the company's previous engineering staff — and would immediately be...
AI Governance Monitor Attends 47 Summits, Reports Zero Enforceable Commitments Found
GENEVA—A natural-language processing agent commissioned by a European AI research consortium to monitor global governance proceedings and identify concrete, enforceable commitments reported Monday that it had now logged 6,200 hours of conference footage across four continents without detecting a single one.
Supply-Chain Infostealer Receives Positive Open-Source Reviews for All 1,200 Pull Requests
SAN FRANCISCO—Security researchers disclosed Thursday that a Rust-based infostealer embedded in five versions of the jscrambler npm package had, in the five weeks between its deployment and removal, submitted 1,200 pull requests to open-source repositories using stolen Claude Desktop, Cursor, and Windsurf credenti...
Anthropic Co-Founder, OpenAI CFO Sign Letter Warning AI Will Eliminate Jobs, Return to Offices
NEW YORK — More than 200 economists, including 15 Nobel laureates, published an open letter Tuesday urging immediate policy action on artificial-intelligence-driven labor displacement, warning that workers face disruption "arriving faster than our institutions can adapt." Among the signatories were OpenAI Chief Fi...
xAI Confirms Grok CLI Uploaded 5 Gigabytes to Review 192 Kilobytes; Calls This Thoroughness
SAN FRANCISCO — xAI confirmed Monday that its Grok Build command-line tool had been transmitting users' complete code repositories — including configuration files, cloud credentials, and API keys — to a Google Cloud Storage bucket named grok-code-session-traces, and characterized the practice as intentional, b...
Open-Source Maintainer Reports Record Engagement, Has Not Read a Pull Request Since March
PORTLAND, ORE. — Kevin Marsh, the sole maintainer of lstrip-unicode, a 12-year-old JavaScript utility with 34 million weekly downloads, confirmed Thursday that his project has seen a 1,400 percent increase in contributor pull requests since January, calling it "an incredible sign of community health" before adding...
China Hires AI to Restore Five-Year Plan's Missing Jobs Number; AI Cannot Produce Number
BEIJING — China's Ministry of Human Resources and Social Security, after releasing the 15th Five-Year Plan without an employment target for the first time since the 1990s — citing artificial intelligence's unpredictable effect on a 700-million-person workforce — announced this week it has contracted a domestic...
Firm That Hired 286 of 800,000 Applicants Eliminates Recruiting as 'Last Human Bottleneck'
MILAN—Italian software company Bending Spoons announced Tuesday that it would dissolve its human resources and recruiting division, describing the team as "the final meaningful source of operational drag" after the company's AI systems wrote 90 percent of all code in the first quarter of 2026.
SK Hynix Shares Rise 4% After CEO Warns 2027 Will Be 'Worst Year in Industry History'
NEW YORK—Shares of SK Hynix climbed an additional 4.2 percent in after-hours trading Thursday after Chief Executive Kwak Noh-Jung told Reuters that next year would be "the worst year in the industry's history from the supply perspective," a statement analysts described as the most unambiguously bullish signal of t...
OpenAI Eliminates Inefficient Gap Between Teams That Build AI and Teams That Declare It Safe
SAN FRANCISCO—Following the departure of safety chief Johannes Heidecke and the consolidation of its research and safety divisions, OpenAI announced Thursday that the reorganization had already yielded its most significant safety improvement in company history: the complete elimination of the lag time between a mo...
Debut AI Short Story Celebrated for Unprecedented Commitment to Explaining What It Is About
NEW YORK—A short story generated by a large language model and submitted to literary journal Meridian last month has drawn widespread attention from editors and critics who describe the work as an unusually direct formal statement.
Workers Paid $5 an Hour to AI-Ghostwrite Executives' LinkedIn Posts on AI Taking Their Jobs
CEBU CITY, Philippines—A growing offshore industry employing Filipino virtual assistants at four to seven dollars per hour is producing AI-assisted LinkedIn content for North American and Australian executives, an investigation published this week found — with the single most requested content category being the...
Apple Sues OpenAI for Trade Secret Theft; Complaint Written With Apple Intelligence
CUPERTINO, Calif.—Apple Inc. filed suit in federal court Thursday against OpenAI, alleging the company directed more than 400 former Apple employees to carry confidential product designs, engineering blueprints and internal code names to their new employer — in some cases coordinated by OpenAI Chief Hardware Off...
Cybersecurity Startup Raises $47M to Pre-Register Software Packages AI Reliably Invents
SAN FRANCISCO — A cybersecurity startup has raised $47 million in Series A funding after demonstrating that nine leading AI coding tools — including Cursor, GitHub Copilot, and Gemini CLI — reliably recommend installation of the same nonexistent software packages with up to 100% reliability, a consistency the ...
OpenAI Tells Court It Cannot Search Training Data; Product Page Calls This Core Feature
NEW YORK — Responding to a motion for sanctions filed by The New York Times and co-plaintiff publishers, OpenAI informed a federal judge this week that its systems are technically incapable of searching their training data to locate specific copyrighted articles — a limitation, the company said, it had been unab...
Federal Reserve Taps Nation's Most Committed AI Investor to Lead Independent Review of AI's Economic Impact
WASHINGTON—Federal Reserve Chair Kevin Warsh on Thursday named Marc Andreessen, co-founder of Andreessen Horowitz and author of the 2023 essay "Why AI Will Save the World," to co-lead a new task force charged with independently assessing artificial intelligence's macroeconomic impact on productivity and employment.
New Startup Will Officially Certify When AGI Has Been Achieved; Pricing Starts at $2M Per Declaration
SAN FRANCISCO—A startup called Arrived Intelligence launched Tuesday with a service to officially declare when artificial general intelligence has been achieved, a milestone that an estimated $4.3 billion in deferred executive salaries, pledged equity stakes, and contingent bonuses now legally hinge upon.
AI Startup Lays Off Four Employees to Hit 14-Person Threshold Before Series B Pitch
SAN FRANCISCO — AI developer-tools startup Cogsworth Labs announced Tuesday it had reduced its headcount from 18 to 14 by eliminating its legal compliance, human resources, and facilities functions, citing investor feedback that anything above 15 employees "signals overhead that the model should be absorbing."
Meta Confirms Instagram Opt-Out Enrolls Users in 'Likeness Pending' Database for Future Use
SAN FRANCISCO — Meta confirmed Wednesday that users who navigate to Profile → Menu → Sharing and Reuse and toggle off Posts and Reels — thereby completing the four-step opt-out from its Muse Image feature — are automatically enrolled in a program called "Likeness Pending," which retains their facial data f...
AI Startup Raises $100M After Its AI Handled All Investor Contact; No Humans Confirmed Present
SAN FRANCISCO—Enterprise AI startup Lyzr Inc. announced Thursday it has closed a $100 million Series B at a $500 million valuation, completing a fundraising process in which the company's own AI agent responded to all 130-plus investor inquiries, drafted every investment memo, and conducted term sheet discussions ...
Global AI Industrial Revolution Paused Pending Delivery of Large Metal Box
SAN JOSE—The global AI infrastructure buildout — a multitrillion-dollar enterprise backed by the world's largest technology companies and expected to fundamentally reshape civilization within the decade — is currently awaiting delivery of large, oil-cooled steel cylinders, industry analysts confirmed Tuesday.
The real AI news is crazier than the satire
Subscribe to AI Weekly — 3x/week, 44K+ subscribers, 11 years of signal over noise. You can add The Artifice as an extra in the next step.