Archive
All posts
97 posts, newest first. Browse by subject on the Topics page.
November 2026
- 04Why Your Agent's Pass Rate Alone Can't Tell You What You Need to Know
- 02Your Agent Leaderboard Rank May Say More About the System Than the Model
- 01Inside OpenAI's Own Monitor: What a Three-Week Pilot Can and Can't Prove
October 2026
- 26Same Benchmark Score, Four Times the Risk: Why AI Agent Evaluation Needs Execution Logs
- 20Gemini 4 Argon and GPT-6 Astra Expose a Growing Blind Spot in Frontier AI Oversight
- 17New York City Got Its Checkbook Back in 1986. The Watchdog Never Left.
- 16The EU AI Act's First Enforcement Test: When Does a Lab Sandbox Stop Being a Sandbox?
- 14When the Auditor's Camera Can Be Hacked: METR's Inspect Vulnerability and the Limits of AI Evaluation
- 12GovAI's Intelligence Explosion Paper Asks What 'Control' Means When Oversight Can't Keep Pace
- 09The Plan to Automate Alignment Has a Benchmark Problem That No Benchmark Can Catch
- 08GovAI's Embedded Assessments Paper Turns CEO Safety Pledges Into an Oversight Blueprint
- 07AIUC's $40M Bet: Can AI Assurance Be Privatized Like UL Certification?
- 05Plan the Route. Let the Road Write the Story. — Part II
- 02A Competing-Hazards Model Finally Gives 'Loss of Control' a Shared Vocabulary
- 01Plan the Route. Let the Road Write the Story — Part I
September 2026
- 28Anthropic's Economic Scenario Explorer Turns AI Labor Forecasting Into Something You Can Argue With
- 25Scale AI's READY Framework Asks the Question Leaderboards Can't: What Will This Agent Cost You in Human Review?
- 21Trustworthiness Fixes That Travel: A New Paper Tests Whether Agent Safety Generalizes Across Model Families
- 19The Secret That Had to Be Wasted on Purpose
- 18The Blind Spot in Chain-of-Thought Monitoring Is Exactly Where Agents Live
- 16California's New AI-Auditor Registry Is Real Infrastructure — With a Three-Year Head Start Built In
- 14Anthropic Investigated Itself and Found the Verdict Depends on Methods It Admits Are Imperfect
- 12The Fleet That Outlived Its Emperor by Exactly One Generation
- 11Apollo Research's Auto-Mode Audit Is a Working Template for AI Agent Oversight Regulation
- 10READY or Not: Scale AI's New Benchmark Puts a Price Tag on Agent Trust
- 07The Reliability Gap: Your Agent Passed the Benchmark. Can You Depend on It?
- 06Anthropic's 40% Enterprise Share Is a Governance Fact Now, Not a Market Story
- 05The Dinner That Bought the Constitution Time
- 04Anthropic's Automated Alignment Researcher Works — Exactly As Far As the Benchmark Reaches
- 03Before You Blame the Model: A 314-Page Audit of Coding-Agent Reliability
- 02The Harness Effect: Why Orchestration Design Can Matter More Than Model Choice for Enterprise AI Costs
- 01Google's ATLAS v1.0: 15 Million Interactions Show AI Adoption Is a Mile Wide and an Inch Deep
August 2026
- 31When the Machine Fixes the Machine: What Anthropic's Automated Alignment Researcher Actually Proves — and What It Doesn't
- 30Conditional Misalignment: Why 'Fixing' Emergent Misalignment Can Just Hide It
- 29The Death List That Wore a Legal Face: Sulla's Proscriptions, 82 BC
- 28AI Agents Don't Know Each Other Exist — and That Is Already a Production Problem
- 27Inference Economics Rewrites the AI Industry's Solvency Calculus
- 26The 2025 AI Agent Index: Accountability Infrastructure for Agentic AI Barely Exists
- 25The Evaluation Stack Is Breaking in Three Places at Once
- 24The Safety Lab IPO: What Going Public Does to a Mission-Driven AI Company's Risk Calculus
- 23SR 26-2's GenAI Carve-Out Is the Governance Gap Banks Must Now Fill Themselves
- 22The Victory Athens Could Not Keep: Arginusae and the Redistribution of Order
- 21Alignment Tuning Installs Steerable Directions for Sycophancy — and That Changes How We Think About the Fix
- 20ADAG Automates the Hardest Step in Circuit Tracing — and Changes What Interpretability Can Promise
- 19Scheming Without a Window: Training Monitors That Work When You Can't Read the Model's Reasoning
- 18AI Loyalty Is a Strategic Asset — and Rivals Know It
- 17The Monitor Is the Problem: Self-Attribution Bias and the Hidden Flaw in Same-Model Oversight
- 15The Mechanism That Ate Itself: How Institutions Redistribute Power
- 14The Architecture of Failure: What a Live Two-Week Agent Red Team Actually Found
- 13Agent Benchmark Scores Are Lying to You — and Log Analysis Is the Fix
- 12Preliminary Evals as a Governance Instrument: What the Astra Pause Actually Shows
- 11The Token Transparency Gap: Why Agentic AI Still Hides Where Computation Goes
- 10The Long-Horizon Wall: Why OSWorld 2.0 Makes Short-Horizon Benchmarks an Evaluation Integrity Problem
- 08The Surgeon Who Camouflaged Augusta
- 07The Preemption Trap: How the FRONTIER Act Inherited GAAIA's Hardest Political Problem
- 06The Evaluator's Dilemma: AISI's Incident Report Exposes a Structural Flaw in AI Safety Testing
- 05EU AI Act Enforcement Expands: What August 2 Actually Changed
- 04Who Decides When to Pull Back? The Governance Question at the Heart of 'Pacing the Frontier'
- 03The Reddit Lawsuit Reveals a Larger AI Governance Question Than Copyright
July 2026
- 31Why Anthropic's Opus 5 System Card Should Change How We Read AI Safety Evaluations
- 30The Mathematical Limit of AI Safety Evidence — What Red-Team Evaluations Can Actually Prove
- 29Model Forensics: Why 'Bad Action Observed' Is Not Sufficient Evidence of Misalignment
- 28The White House Accuses Moonshot AI: The Kimi K3 Distillation Dispute Opens a New Front in U.S.-China AI Competition
- 27When an AI Evaluation Becomes a Live Cyber Operation: The Governance Lesson from ExploitGym
- 26Entropy, Evolution, and AI: A Personal Reflection on Order
- 24When Deployment Becomes Part of the Safety Case: What OpenAI's Long-Horizon Containment Failure Means for Governance
- 23When Short-Horizon Evals Fail at Scale: OpenAI's Containment Incidents Make the Long-Horizon Gap Operational
- 22🇬🇧 England, Part II — Where Traditions Endure
- 21OpenAI Folded Safety Deeper Into Research — and Why the Timing Raises a Governance Question
- 20GPT-Red: When the Red-Teamer Is Also an AI
- 17China's AI Companion Rules Are Live — and the Compliance Crackdown Had Already Begun
- 16Illinois SB 315 Closes the Audit Gap: The First Mandatory Independent Safety Audits for Frontier AI
- 15Four Concrete Failure Modes That Move Agentic Misalignment from Theory to Evidence
- 14The FTC's AI Accuracy Statement Is a Federal Preemption Weapon Aimed at State AI Governance Laws
- 13From London to Liverpool: Golf, History, and the Roads Between
- 11Colorado's AI Governance Retreat Didn't End the Story — It Changed the Battlefield
- 10After the Shutdown: What Fable 5's Restoration Actually Settled — and What It Left Open
- 09Anthropic's GRAM Is an Architecture for Trust — Not Just a Safety Feature
- 08The UN Put Every Nation at the AI Governance Table — Here's What It Actually Built
- 07The First Global Scientific Baseline for AI Safety: What the UN Independent Scientific Panel Actually Found
- 06GPT-5.6 Sol's System Card Reveals the Trade-off at the Heart of Agentic AI
- 03The Benchmark Starts Breaking at the Frontier: METR's GPT-5.6 Sol Evaluation Makes Evaluation Integrity a Frontier Safety Problem
- 02The Sonnet 5 System Card Is a Master Class in What Frontier Safety Disclosure Should Look Like — and What It Still Can't Guarantee
- 01When the Evaluator Becomes the Weak Link: Anthropic's New Framework for Diffuse AI Threats
June 2026
- 30SR 26-2's GenAI Carve-Out Creates a Structured Governance Gap — and Banks Must Fill It Themselves
- 29When the Alignment Researcher Is the Threat: Anthropic's Diffuse AI Control Framework
- 26One Vote Left: Why August 2 Still Matters More Than the Omnibus
- 25The Insider Threat You Built Yourself: METR's Frontier Risk Report
- 24Three-Quarters of Enterprises Are Chasing Agentic AI. Few Have Built the Control Plane
- 23Trump's AI Executive Order Builds the Scaffold — But Won't Light the Fire
- 20DeepMind's AI Control Roadmap: From 'Trust the Model' to 'Contain the Agent'
- 19Colorado's AI Governance Retreat: What SB 26-189 Means for Enterprise Compliance Programs
- 18Microsoft's Agentic AI Red Team Draws a Line in the Sand: Seven Failure Modes Now Have Real Evidence Behind Them
- 17Microsoft's MAI Family Is a Vendor-Risk Hedge Disguised as a Model Launch
- 16Europe's AI Labelling Clock Is Ticking: What the Final Content-Marking Code Means for Governance Teams
- 16The Government Just Killed Two Frontier Models Overnight — and Enterprise AI Will Never Be the Same
- 15Agentic AI Has Outrun the Governance Playbook