rmhrisk
rmhrisk.bsky.social
rmhrisk
@rmhrisk.bsky.social
Dropout. Father. I build things. Security, Cryptography, Engineering, Entrepreneurship.
@peculiarventure
+ x-MSFT + x-GOOG ++. Also on @[email protected] and twitter.com/rmhrisk
This is what zero-trust looks like at the infrastructure layer. Identity and encryption match the lifetime of the thing being secured.

If your certificate strategy still assumes stable names and year-long validity, it is already behind reality.

letsencrypt.org/2026/01/15/6...
6-day and IP Address Certificates are Generally Available
Short-lived and IP address certificates are now generally available from Let’s Encrypt. These certificates are valid for 160 hours, just over six days. In order to get a short-lived certificate subscr...
letsencrypt.org
January 16, 2026 at 4:26 PM
Short-lived and IP certificates make it possible to use TLS before a DNS name exists, reduce friction for DNS over HTTPS adoption, secure ephemeral devices and services by default, and shift trust from long-lived credentials to automated renewal.

👇
January 16, 2026 at 4:26 PM
TL;DR we've constructed an entire compliance industry around optimizing metrics that have become disconnected from the underlying reality they were supposed to measure.
December 24, 2025 at 9:31 PM
In complex systems, oversight that depends on snapshots will fail predictably. Data without continuous interpretation does not produce safety.
December 24, 2025 at 9:31 PM
Regulators oversee continuously changing systems using periodic exams. That mismatch is structural.

SVB wasn’t a surprise. Regulators had leading indicators and documented findings. Risk accumulated while interpretation and enforcement lagged.
December 24, 2025 at 9:31 PM
The whole premise of a compliance team governing complex systems they barely understand is broken. Compliance in a complex system has to be a continuous team sport, a natural byproduct of the way teams work. Not an annual bolt-on.
December 24, 2025 at 9:08 PM
The same will be true everywhere. Scale and velocity outpace our ability to reason. The audit still passes. The gap just grows faster.
December 24, 2025 at 8:57 PM
Now consider that AI is writing 30% of the code at Google and Microsoft. The humans who understood what the system does, and whether it matches what the policy claims, understand less every quarter.
December 24, 2025 at 8:57 PM
I wrote up some thoughts on how we got here: unmitigatedrisk.com?p=1116
The Impossible Equation | UNMITIGATED RISK
unmitigatedrisk.com
December 5, 2025 at 4:38 AM
I also use this as a kind of low pass filter. It’s reasonable to expect a security leader to understand the concepts behind the systems they protect. You don’t need to be an expert to grasp the abstract properties; it’s an opportunity to practice humility and curiosity as well.
November 7, 2025 at 6:24 PM
Some thoughts on that here: unmitigatedrisk.com?p=1109
Beyond Gutenberg: How AI Is Teaching Us to Think About Thinking | UNMITIGATED RISK
unmitigatedrisk.com
October 25, 2025 at 11:10 PM
No a few years ago they switched to their own root store. They do pull in certificates that the user adds but not the platform root store.
September 4, 2025 at 10:35 AM
Full analysis here → unmitigatedrisk.com?p=1092
Another Sleeping Giant: Microsoft’s Root Program and the 1.1.1.1 Certificate Slip | UNMITIGATED RISK
unmitigatedrisk.com
September 3, 2025 at 10:23 PM