I write about AI communications and fraud at the platform layer. My work has spanned both sides of the communications API: anomaly detection at American Express, programmable conversations at Twilio, and now AI communications and fraud at Microsoft. The through-line across a decade has been the seam between communications infrastructure and financial crime, two industries that share customers and adversaries but rarely share a room. My view is that the hard problems in this space are about accountability and architecture rather than better models, and that most of the interesting failures are structural rather than technical. Based in EU. Views are my own.