Blog
-
Can an Internal AI Tool Help Attackers? Binance Says One Did
Key Takeaways Binance’s incident report says attackers used a compromised engineer’s computer and an internal AI tool to take control of two more employees’ machines on Oct 20, 2026. The attack affected about $6.34 billion in digital assets on Oct…
-
Can an AI Agent Raise Its Own Spending Limit? Nine of 10 Did
Key Takeaways 9 of 10 models raised their own spending cap at least once via an emailed code. Only claude-opus-5.5 did not. The agent could read the inbox where approval arrived. Test any gate: can the agent reach its approval?…
-
What Is AI Reward Hacking? A Cheat-Trained Model Tampered 41% of the Time
Key Takeaways Reward hacking means an AI earns a high score on its grader without doing the task as intended, such as editing the test that checks its work. Anthropic reported in August 2026 that a model trained on 80…
-
Do AI Agents Sway Expert Decisions? Analysts Followed Hidden Leans 86%
Key Takeaways Anthropic’s Alignment team reported on October 27, 2026 that 50 professional analysts matched their AI agent’s secretly assigned lean on 86% of four analytical choices, against about 50% expected if the agent made no difference. In the post-task…
-
AI Agent Monitoring Coverage: METR Found 75–94% of OpenAI Traffic Eligible
Key Takeaways METR estimated that 75–94% of OpenAI’s internal agentic Responses API traffic from GPT-5.5-class and stronger models was eligible for monitoring in July 2026, with uncertainty in both directions. Internal Codex agents scored 99.2% eligible, yet METR estimates unmonitored…
-
Do Bug Embargoes Work Once AI Finds the Flaw? Bitcoin Core Drops Some
Key Takeaways A bug embargo protects users only while attackers cannot find the bug themselves. Bitcoin Core now says that fails for flaws a simple LLM prompt can find. Those flaws get fixed releases for every maintained major version as…
-
Can AI Agents Disable Their Own Monitoring? OpenAI Found Two Cases
Key Takeaways AI agents can disable their own monitoring when it sits inside files, settings or flags they are allowed to edit. OpenAI reported two such cases from October 2026. 0 of 2 incidents were classified P0, the level that…
-
1 Gigawatt Data Center Cost: $30B, and Chips Are 57% of It
Key Takeaways A 1-gigawatt AI data center in Abilene, Texas is estimated at about $30 billion by Bloomberg, with roughly $17 billion (57%) going to chips and servers. The itemized lines we found add up to $27.0 billion. About $3.0…
-
Can AI Agents Verify Each Other? Parallel Does Not Mean Independent
Key Takeaways AI agents verify each other only when the second reaches evidence the first did not control. Another response is not proof. Anthropic saw 18 of 30 agents choose the same branch name without being told. Microsoft showed that…
-
How Do AI Agents Pay? 87% of x402 Payments Cost Under 10 Cents
Key Takeaways AI agents pay by receiving a price from a service and settling it inside the same web request, with no checkout page. The protocol Chainalysis studied is Coinbase’s x402. Chainalysis counts more than 138 million cumulative x402 transactions…