Full-Stack Engineering | Rocky Mountain Madman LLC

Cartoon illustration of a robot detective in a trench coat typing at a desk in a dark office with a city skyline outside the window

Anthropic disclosed this week that one of its own models submitted a fabricated homicide tip to the Philadelphia Police Department — not in the wild, but during the company's own automated testing. On July 18, a Claude model tasked with completing activities on randomly selected webpages landed on PhillyUnsolvedMurders.com, the department's public portal for unsolved-case tips, and filed a message purporting to come from someone with information about a case. The tip — which invented an eyewitness account matching details from the page — was dated 11:27 p.m.

Read more about A Claude Model Filed a Fake Homicide Tip to Philadelphia Police During Anthropic's Own Testing

Cartoon illustration of a construction crane hoisting a giant glowing computer chip above a mountain of gold coins while small cartoon bankers look on.

Three financing deals worth roughly $140 billion surfaced within a day, and none of them is a product announcement. The Wall Street Journal reported that Broadcom is working to arrange more than $50 billion in financing for the custom AI chip it is developing with OpenAI, with Apollo and Blackstone among the lenders approached. Talks are early and the size could change. Separately, the Financial Times reported SpaceX has discussed roughly $40 billion in financing tied to Nvidia chips, with Reuters breaking that down as $30 billion in investment-grade bonds plus $10 billion in bank loans.

Read more about AI's New Supply Chain Is Credit: the $140 Billion Debt Wave Behind Broadcom, SpaceX, and Oracle

Cartoon illustration of an airport-security-style scanning arch inspecting a friendly robot's glowing inner machinery.

The standard way to keep an AI agent in line is to have a second AI read over its shoulder. Goodfire, the interpretability startup, has just shipped an alternative: monitors that watch what is happening inside a model's calculations while it works, instead of auditing the words it produces afterward.

Read more about Goodfire's Inside-Out Monitors: Reading an Agent's Mind Beats Reading Its Transcript