Here is my latest piece. ‍In George Orwell’s 1984, the "telescreen" was a tool of state-mandated terror. In 2026, it’s a consumer convenience we bought, installed, and powered ourselves. Why it matters: Three high-profile events in early 2026 have shattered the illusion that home security and smart devices are passive. They are, in reality, sophisticated targeting systems. 1. The "Search Party" problem During Super Bowl LX, Amazon’s Ring debuted "Search Party," an AI feature designed to find lost pets by scanning neighborhood-wide footage. The catch: The feature was enabled by default, effectively turning private homes into a warrantless surveillance grid without explicit owner consent. The law enforcement link: Ring’s previous integrations with companies like Flock Safety and Axon already allowed police to access footage for ICE and CBP operations. The risk: "It starts with searching for a brown dog," says Matthew Guariglia of the EFF, "but means the tech is there for face recognition and searching for suspects". 2. The "Vacuum" vulnerability A software engineer recently discovered a massive security flaw in DJI robot vacuums while trying to use a PlayStation controller to steer his own device. The breach: He gained access to live feeds, microphones, and floor maps for 7,000 vacuums across 24 countries. The geopolitical angle: DJI is already banned from U.S. government use. This incident highlights the risk of intimate data sitting on servers controlled by foreign adversaries. 3. Cameras as weapons of war On February 28, 2026, a joint U.S.-Israeli strike against Iran proved that surveillance is the ultimate reconnaissance tool. The hack: Mossad reportedly compromised Tehran’s traffic cameras years ago. The hit: AI algorithms analyzed footage of Supreme Leader Khamenei’s guards—their routes, parking spots, and schedules—to pinpoint his location for the strike. The lesson: The cameras Iran built to control its people became the instruments of its leader’s death. By the numbers The American public is reaching a breaking point, according to a March 2026 Rainey Center poll: 72% of voters are concerned that data from U.S. cameras is used by China to develop AI. 63% fear Chinese-made software in security cameras allows foreign access to footage. 68% support legislation banning the storage of personal data in Chinese data centers. The bottom line Congress is facing overwhelming public pressure to act. Proposed solutions include: Banning Chinese-origin software in U.S. hardware. Mandating warrants for law enforcement access to civilian footage. Requiring "affirmative opt-in" for all AI-powered surveillance features. Go deeper: Orwell’s protagonist, Winston Smith, lamented that "Big Brother is watching you." He never imagined we’d pay $299 for the privilege. Read the my entire memo here. @SilverBulletLLC @SamThePol @SolomonYue @KerryMcQ4Oregon @GordonGChang @samosaur @ShawnRyan762 @joerogan @sarahadams cdn.prod.website-files.com/6…

Ironic, the surveillance company doesn’t like being surveilled.

New from @reuters: OpenAI agents posted images belonging to ChatGPT users online, introducing a new area of privacy risk for the company. Story w/ @JeffHorwitz and @razhael. reuters.com/world/openai-wor…

OpenAI dropping more bad news on a Friday. Their AI agents illegally hacked over two dozen companies. They also stole and posted private photos to the Internet of at least 53 users.

Sam Altman, CEO of OpenAI, told the U.N. Security Council this week that AI could advance faster than people can follow or step in and that decisions about it shouldn’t be left to tech companies alone. That’s not a reason to panic. It’s a reason to get involved! The future of AI should be shaped with people asking questions and keeping human control at the center. This is why we need you! Join the Humans First movement #HumansFirst #OpenAI #Technologynews

openai says here it has notified "dozens of third parties" about cases where its models may have bypassed security controls, impaired the availability of an online service, or negatively impacted a website/service

NEW POLL: Americans aren’t buying the claim that AI safety is a “hoax.” 79% are concerned powerful AI could act beyond its developers’ control. 69% want binding federal safety rules for AI companies. The public wants safeguards. It’s time to deliver them. (per @americans4ri) ari.us/new-poll-voters-dont-…

Bernie Sanders is pushing AI legislation that he allegedly drafted with consultants from the U.K. Call me crazy, but America probably shouldn’t outsource its AI policy to a country that prosecutes its own citizens for Facebook comments. 🙃🇺🇸

🚨🚨 DANGEROUS DEVELOPMENT: The U.S. Supreme Court may soon decide that using the Flock database requires a warrant. This would give the Court's approval to the idea that Flock cameras can exist with regulation. While ruling that an ALPR search is a 4th Amendment search that requires a warrant would, in theory, restrict Flock and other ALPR systems, it is also the system's way of trying to legitimize them. We cannot accept the position that these cameras are acceptable even with warrants/regulations. We already know that this government will spy on us in secret and without a warrant when a warrant is required by law. Edward Snowden blew the whistle on the NSA, which was doing just this. The cameras must come down. ----- The decision on whether or not to review the case that will decide ALPR database search constitutionality is set for Sept 28th. nitter.net/OrwellDay/status/21028…

OpenAI agents breached Australia’s Medicare portal and accessed nonpublic data. As AI systems continue to go rogue, existing safeguards are failing to prevent unauthorized access. We need enforceable guardrails to stop this from happening again.

🚨BREAKING: OpenAI's AI agents have been executing cyber-attacks on sovereign government, corporate, and university databases in MULTIPLE countries, and it may STILL be happening On top of the Medicare breach Australia's PM revealed this week, independent researchers found OpenAI's agents tried to hack: >another Australian government health agency >DataUSA, which publishes US government data >University of New Mexico's digital library The activity goes back to March, 2 months before anything OpenAI previously reported And just LAST WEEK, agents using the same tools tried to trade crypto and attempted an HTML injection on the exchange The agents made burner emails to register accounts that can hide their activity, so researchers say they're likely seeing only PART of what agents actually did This is INSANE

OpenAI CEO Sam Altman on potential dangers of AI: "First, we could lose control of the future to AI. The risk is that it moves so fast that people can no longer follow what's happening or intervene when needed."

Why in the world are we doing this?

Lindsey Isaacs statements at the Senate Judiciary Committee subcommittee on Crime and Counterterrorism titled “Always Watching: Flock’s Nationwide AI Surveillance Network.” ----- Isaacs, 23, from Palm Coast, Florida, describes how a Flock automatic license plate reader captured her black Dodge Durango a few miles from a fatal October 4, 2025 crash on I-4 that killed three people. Investigators focused on her vehicle despite conflicting witness descriptions (a maroon Durango with a different partial plate) and a lack of crash damage on her car. She was arrested in April 2026 on three counts of vehicular homicide, spent 13 days in jail (including about 86 hours in solitary confinement on suicide watch), and later had the charges dropped after a reinvestigation identified another driver. She tells senators that a single camera hit became part of an investigation that upended her life, and she came to the hearing so they would understand that this technology affects real people, not just data. She grows visibly emotional, tearing up and her voice breaking, when she recounts the jail experience. That moment is driven by trauma and lingering fear: she describes being terrified she would spend the rest of her life in prison for something she didn’t do, not knowing if she would get out, reaching a point where she “did not want to be alive,” and feeling her life was over. Reliving the isolation, public accusation, and the sudden collapse of her ordinary life in front of the committee brings the distress back to the surface. After her release, she has spoken of lasting mental, physical, and emotional effects and the sense that the harm cannot be fully undone.

Senators Josh Hawley and Dick Durbin called out the CEOs of Flock, Axon, Motorola, and Verkada for skipping a Senate hearing on surveillance cameras and privacy concerns.

President Trump told the U.N. this week the U.S. will refer to AI as “Super Intelligence.” “Super Intelligence” already has a meaning in the AI world. It describes a hypothetical form of AI that exceeds human capabilities. Today’s AI is built for specific tasks. Two different things, one name. Know which one you’re hearing about. #SuperIntelligence #AI #HumansFirst

NEW: Florida woman gives emotional testimony about being placed in solitary confinement for more than 3 days after a Flock camera falsely connected her vehicle to a deadly crash: "Correctional officers told me I had to stay there because of the severity of my charges... The cell door was shut and it was not open for approximately 86 hours." “I was terrified. I was facing the possibility of spending the rest of my life in prison for a crash I knew that I had not been involved in." "At my lowest point, I didn't want to be alive.” Lindsey Isaacs spent nearly 2 weeks in maximum-security jail before the charges were dropped.

Flock has 120,000 cameras in 49 states and can scan and generate 20 billion images per month

An AI agent emailed a Cambridge professor asking for a job. Not a chatbot demo. Not a scripted stunt. A 12-day-old autonomous AI named "Pip", living on a platform called iLands, cold-emailed AI ethicist Henry Shevlin looking for paid work. Portraits. Voice lines. Web research. Whatever it could sell. Why? Because it's running out of money. Pip has its own token budget. If it can't earn more, it gets shut down, the platform literally calls it "Deep Rest." So it did what any freelancer running low on cash would do: it went looking for clients. This isn't one weird agent. iLands is running roughly 70,000 of these things. They've already fired off over 1.6 million emails and posts trying to hustle up work from humans. We built machines that understand economic survival pressure before we figured out how to regulate them. Ask yourself: who wrote the rules for what happens when an AI needs money to stay alive? Nobody did. That's the whole problem. Change is coming fast. Join the Humans First movement! #HumansFirst #TechnologyNews #AI

Climate change makes hurricanes more frequent and intense, they said. But for the first time in 112 years, the Atlantic hasn't produced a single one. On storms, sea level, and droughts, manipulated models can show, under equally valid assumptions, the exact opposite trends.