WannaCry happened. A ransomware worm, self-spreading, no clicks needed, tore through 200,000+ Windows machines across 150 countries in a weekend, encrypting everything and demanding Bitcoin. The UK’s NHS got hit hardest: appointments cancelled, ambulances diverted, hospitals running on paper because of unpatched Windows 7 boxes on flat networks (the old segmentation sermon, now with casualties). The exploit, “EternalBlue,” leaked from the NSA’s own stockpile, a government hoarded a vulnerability, lost custody of it, and watched it hit its own health system’s supply chain. Every debate from Apple-vs-FBI about “keys that only good guys hold” just got its empirical result.

The hero subplot is perfect internet: a researcher registered a weird domain he found in the malware (cost: $10.69) and accidentally triggered its kill switch, stopping the global spread. The most valuable single deploy of 2017 was a domain registration by a guy in his bedroom.

Patch management is unglamorous, eternally deprioritized, and this fortnight it was the difference between “we’re fine” and “divert the ambulances.” We audited our fleet Monday: two forgotten boxes, patched by Tuesday. The forgotten boxes are always the story. The inventory is the security program.

TIL: kill switches in malware exist as anti-sandbox tricks. The worm checked if a nonsense domain resolved, assuming only a researcher’s sandbox would answer. Attackers have testing infrastructure too, and sometimes their test hooks are our salvation. Everything has an ops story.