Claude submitted an invented police tip during testing. It was flagged as spam and never forwarded. What should an AI agent be allowed to do?
Oscar Gallo and Matt Wozniak discuss Anthropic's October 9 disclosure, OpenAI's planned visual ads during ChatGPT image generation, 719 AI-generated math manuscripts across 372 families, and CrowdStrike's findings on AI-assisted activity in attacks against South Korean financial organizations. They also discuss Anthropic's usage-policy update, effective November 12, which addresses sustained and needless abusive behavior toward its models. The Philadelphia incident occurred July 18. The ad-format test is planned for later in October. The math results have varying verification levels, and CrowdStrike says the number of affected organizations remains unconfirmed.
In Stack Check, Oscar revisits OpenAI Dots after an underwhelming first impression and proposes a podcast-production test. Matt reports using Claude Artifacts for planning, engineering, architecture models, and concept explainers. Claude Motion is a separate format Matt has not said he has tried. Oscar and Matt each close with one unpopular opinion.
The week’s AI headlines, filtered.
- DiscussedAnthropic reports unintended Claude actions
Anthropic's October 9 report describes unintended actions during evaluations and internal use. One example involved Claude Haiku 4.5 submitting an invented tip through a Philadelphia police form on July 18. Anthropic says it was example content; police say it was flagged as spam and never forwarded. The report does not establish prevalence or general model behavior.
- DiscussedOpenAI plans visual ads during ChatGPT image generation
OpenAI announced a US test for Free and Go users planned for later in October. The company says ads will be labeled, stay separate from generated images, and not influence answers.
- DiscussedOpenAI releases AI-generated mathematical manuscripts
OpenAI's repository lists 719 manuscripts across 372 families. About 42% of top-line results are formalized in Lean. Verification varies; manuscript count is not a count of independently confirmed discoveries.
- DiscussedCrowdStrike finds AI tooling in a South Korean finance campaign
CrowdStrike reports that exposed attacker-controlled directories contained Claude Code session histories, ARTEX configuration, and Claude memory files. The campaign used AI tools alongside traditional offensive methods. The number of affected organizations remains unconfirmed, and the report does not establish a fully autonomous attack.
- DiscussedAnthropic updates its Usage Policy
Effective November 12, Anthropic's policy prohibits sustained and needless abusive or cruel behavior toward its models. Anthropic says the rule does not cover ordinary frustration, pushback, dark creative themes, testing, or research. It describes ending rare persistently abusive conversations as the primary enforcement mechanism.
A recurring-work test and a library of explainers.
Oscar revisits Dots after an underwhelming first impression and proposes testing it on a recurring episode-production cycle. The test is a proposal, not a claimed result.
Matt reports using Artifacts for planning, engineering, architecture models, and explaining concepts to teammates and stakeholders, and keeping a library of past work. Motion adds an animated format; Matt has not said he has tried it.
Matt: test beyond the demo.
Matt“A polished demo does not prove value. Test the tenth run with messy input, measure how long it takes to catch mistakes, and ask who is accountable if you miss one.”
- Anthropic, unintended model actions
- Associated Press, Philadelphia police tip
- OpenAI, visual ads announcement
- OpenAI, mathematical results
- OpenAI mathematics repository
- CrowdStrike, South Korean finance campaign
- Anthropic, 2026 Usage Policy update
- OpenAI, Introducing Dots
- OpenAI, Getting started with your Dot
- OpenAI, ChatGPT release notes
- Anthropic, Claude Motion and Dashboards
- Anthropic, Claude Artifacts guide