Synthetic Minds Newsletter

Synthetic Minds | The Next Cyberattack Wants Your Confidence, Not Your Data

The most dangerous hackers are not after your data or your money. They are after confidence. Supermarket freezers warmed across Europe, plasma freezers at France's blood service reset to 110C, AI accounts billed $100,000 in a day. Why chaos is the product, and why any small business can be a target.

Synthetic Minds | AI Is Becoming a Utility: Hard to Stop, Easy to Switch

Intelligence is becoming a utility: hard to stop, easy to switch. Replit moved most of a $100 million contract between two labs, Meta's shares rose after a $25 billion fine, and Character.AI may go dark. Why switching should be cheap, why fines work like tolls, and what workers left behind are owed.

Synthetic Minds | Who Lent Your AI the Right to Act?

Authority has become something you borrow. An OpenAI model paid from a wallet it found, 952,008 US clinicians asked one AI tool about 271 million questions, and 30 days of silence counts as a yes. Why AI agents need provenance and an undo, and why companies that stop forming judgment lose much more.

Synthetic Minds | Who Holds the Lock When Machines Knock?

Every connected product assumes the person holding it can keep it safe. That assumption has expired. An AI agent found a flaw missed for 11 years; a solar cloud ignored its password; smart glasses opened their cameras to apps. Why security is a strategy question, and why quantum's date is not yours.

Synthetic Minds | Why AI Must Earn Its Trust Every Day

Most companies approve an AI system once, like a copier. In Utah, AI must earn 1,250 doctor-approved decisions per medicine; the UK will monitor AI devices for life; an Anthropic agent sent a fake tip to real police. Trust in AI is a running cost. Plus: why repairing beats rebuilding, almost always.

Synthetic Minds | Your Next Customer Is a Machine That Barely Spends

A hundred AI agents ran a town economy and changed three tenths of one percent of the prices. A rail for agent payments has gone live across 1.5 billion accounts. Mark van Rijmenam on how to read a faint signal: do not change course, and do not ignore it, because exploring it is never wasted effort.

Synthetic Minds | Four AI Labs Agreed to Slow Down, But Only on Their Terms

Anthropic has offered outside evaluators a desk, an access badge and the right to publish without permission, and three rival labs endorsed it. Not one has offered anyone the power to halt the work, and the same plan keeps advanced chips from China. Mark van Rijmenam on watching a lab none can stop.

Synthetic Minds | Big Tech Started Building the Grid

A utility affiliate built two gas plants and a battery for Amazon and Alphabet in under five months by skipping the public-necessity review. The bottleneck is not speed but the wrong test, and that proving a build is net positive should buy a developer the fast lane instead.

Synthetic Minds | Healthcare Enters the Second Half of the Chessboard

A mammogram signed off with no radiologist. Every possible mutation predicted in advance. Cells built inside a patient, an organ grown in a pig. Medicine's oldest shortages are becoming manufacturable, and the steps are compounding fast. Prepare for exponential change, then implement it very slowly.

Synthetic Minds | The Immersive Stack Broke Into Pieces Anyone Can Use

Meta spent a decade trying to own the whole immersive stack. Separate companies are assembling it instead: the world model from one, the platform from another, the glasses from a third. That opens the field to anyone who can build, if the pieces can talk to each other.