RA/QA: “The patient data never leaves the device.” Would you sign off on that claim? Moving clinical AI onto the device has real advantages. Less dependence on connectivity. Faster responses. Less patient data sent to cloud services. But follow the data one step further: patient → prompt → processor → output. The processor has to use the sensitive information. Its physical activity can reveal information about what it is doing. The research I discuss in this week’s essay demonstrates that possibility under controlled laboratory conditions, with a probe close to the processor. That does not establish an attack from a hospital hallway. It does give RA/QA a new question

Local doesn’t mean private Side-channel attacks on edge AI in medical devices. I was a technician in a communications battalion in the Israeli Army. We provided communications infrastructure for an armored division. Two of us handled encryption of trunk communications. Our equipment sat in an armored van and the lines in and out were physical cabling. I was trained as a physicist. Sitting in the van - I often wondered how much of the radiation leaking from those cables could be remotely detected, and at what distances. Today - I got an answer to my old question. Researchers at University of Amsterdam recently demonstrated a proof of concept attack extracting output tokens and user prompts via electromagnetic emissions. See this article on Arxiv published Oct 9, 2026. The deployment of large language models (LLMs) on edge devices is largely motivated by privacy concerns over sharing sensitive user data with cloud servers. The authors demonstrate that keeping inference local does not eliminate privacy risks They put a near-field probe next to the processor and measured what the chip radiated while it was computing.

Can you ignore $27M exposure for patient Mammography data disclosure? I'm reading Walter Isaacson's biography of Elon Musk (stole it from one of my sons). Musk has an insane appetite for risk. And an insane ability to understand details. It's intuitively obvious that more leverage brings more risk. Your competitive advantage creates value. It also creates exposure. I built OpenCRO - the AI chief risk officer with 1500+ AI-enabled medical devices. OpenCRO's first product is a corpus of 1500+ risk models and competitive advantages. Here's an example: Hologic, Inc. is a global medical technology company headquartered in Marlborough, Massachusetts, that develops diagnostics, medical imaging systems, and surgical products focused on women’s health. Company was delisted from the NASDAQ on April 7, 2026, following its acquisition by Blackstone and TPG at a final valuation of $76.01 per share, representing an enterprise value of up to $18.3 billion Current market cap is $14.7BN OpenCRO estimates SLE cost for Mammography data disclosure of $27M.

Hi Anthony I am a pretty serious Sax Player Recording from our last concert youtu.be/GVuGFj30piI?si=if7X…

Tired and tested: Can AI help you sleep better? Onera SleepMap is an AI-enabled device that offers a cloud-based, automated analysis of PSG signals with high accuracy, unlike competitors relying on third-party software. What is the cost of a single loss event due to bugs in the AI? @Onera Health has competitors like Somnoware - Somnoware Diagnostics, X-trodes - X-trodes Sleep System and Neurobit - Neurobit PSG What is the loss in $ of Physiologic signal artifact misclassification ? What is the loss in $ of Population performance drift ? What is the loss in $ of Respiratory event underdetection ? What is the loss in $ of Respiratory event overcalling

In a 2-10 year march to market how is your software quality? It is rare to see systematic defect reduction projects in MedTech software running in the healthcare provider enterprise, If it were easy, everyone would be doing it. So what makes it so hard? Current development methodologies (including Agile) used by internal development teams are a bad fit for threat analysis of production software systems. The cost of finding and fixing a bug in a production system is regarded as too high. The application developers and IT security teams don’t usually talk to each other. The larger the organization, the more they lose when information gets lost in the cracks. Do your engineering teams talk to your hospital IT teams on a regular basis?

OpenCRO agents will show you a competitive comparison and your top SLE (single loss event) and how much it will cost. DM me if you're interested

Founder content is always about building, winning, raising, scaling, resilience. The West-coast VC narrative is set on Gen Z people. But some founders have aging parents. Aging parents means dependency, decline, responsibility, mortality—and the uncomfortable fact that the supposedly indefatigable founder has a life where they may be the caregiver rather than the person in control. Especially for founders in their 50s and 60s, it's an enormous part of real life and almost absent from the founder conversation. The team deals with aging parents. There is a pale moon over Venice beach. We greet a new anti-design pattern: Solitude. newsletter.dannylieberman.co…

7-step risk analysis loop infographic It's quite remarkable how bad ChatGPT 5.5 is at spelling even with a detailed and very specific prompt