Jan Schaumann

"CrackArmor: Multiple vulnerabilities in AppArmor"

Strong opening by the good folks from Qualys:

"First, we discovered a fundamental vulnerability (a "confused-deputy"
problem) that allows an unprivileged local attacker to load, replace,
and remove arbitrary AppArmor profiles"

https://www.openwall.com/lists/oss-security/2026/03/12/7

Y'all know why casinos use chips instead of cash right? Using "tokens" as your new made-up currency has the same effect of removing the psychological barrier of spending money.

So when AI companies bill you by different types of tokens using astronomical conversion factors, remember that the goal is obfuscation.

What idiot called it “agentic AI” when “Claude Computing” is right there?

(Bonus joke: There is no Claude, only other peoples’ content.)

Tonight seems like a good time to remind folks that the next #nokings protest is planned for March 28th:

https://www.nokings.org/

Mark your calendars, I'll see you there! ✌️

#sotu #fucktrump

Corporate editors now insist on adding AI generated "executive summaries" and "FAQ"s to blog posts, regurgitating the very document they accompany. This is purportedly for SEO (🤮), basically encouraging a "why bother reading the thing" mindset.

Still believing that actual humans are capable of actually reading a text with more than 100 words, I'm rather irritated by that.

I still don't understand where people think Senior Engineers come from if they replace Junior Engineers with AI.

How do they think the next iteration of GIGO regurgitated code will be validated and debugged if nobody understands how the black box works?

Not wanting to understand but just wanting the quick answer is a clear trend and implicit outcome of frequent AI use.

To me, that's the real supply chain risk.

Being forced onto Outlook Office365 with the threat of having to use Outlook or webmail. I don't know how anybody can suffer their email with these interfaces.

So very thankful for the folks who created and documented mutt_oauth2:
https://gitlab.com/muttmua/mutt/-/raw/master/contrib/mutt_oauth2.py.README

Got everything to work using the README and this HOWTO:
https://staff.washington.edu/shrike/unix/configuring-mutt-neomutt-for-imap-access-to-o365-exchange/

And thanks to a kind admin who added a new client-id for mutt, I don't even have to pull the Thunderbird ID to fake that.

Claiming use of drones by drug cartels forces you to close airports seems like a convenient way to reinforce an argument that the US is "at war" with drug cartels, just in case anybody continues to object to the murder of people on boats in the Carribean and Pacific.

Of which there were two more in the last week, but apparently not only does nobody object to them any more, nobody really notices. It's just what we do now.

Karen Hao's "Empire of AI" is really quite good and covers a lot of the megalomaniacal mindsent so pervasive in Silicon Valley circles / AI and the bizarre effective altruism cult. Her conclusion nicely summarizes how OpenAI's "principles" were hollow and trivially redefined as needed by Altman.

A little while ago I was invited to the PQShield podcast "Shielded" to talk about post-quantum cryptography at Akamai. That episode is now available wherever you get your podcasts as well as here:

https://shieldedpodcast.com/e/xn129mv8-compliance-deadlines-customer-reality-and-the-case-for-embedded-tls1-3

Also on youtube:
https://www.youtube.com/watch?v=pcKuJNpb-E0

(I, of course, would rather stick needles into my eyeballs than listen to my own voice or watch myself gesticulating in front of a laptop camera, but ymmv.)

#pqc

If you'd like to have a nice 🤦-y day like me, go ahead, search your internal code repository for the following: curl -k wget --no-check-certificate verify=false InsecureSkipVerify: true ssl_verify_mode: 0 TrustSelfSignedStrategy rejectUnauthorized: false strict-ssl = false http.sslVerify = false --no-verify-ssl --insecure* verify_hostname: 0 in your internal code repository. It'll be great.

Billionaires, man. Greedy motherfuckers. Brin is worth $266B. A one-time 5% billionaire tax (with the money going to healthcare!) would mean he's left with... $242B. Which that year's rate of growth would trivially earn him back. But he just can't. You can do a 99% one-time billionaire tax and he'd be left with *2.6 billion dollars. https://www.nytimes.com/2026/01/28/us/politics/california-billionaires-sergey-brin-campaign.html CA has maybe 200 billionaires in a population of ~40 million. Talk about the richest 0.0005%. Greedy motherfuckers.

Oof, this hits hard.

"AI is Destroying the University and Learning Itself"

https://www.currentaffairs.org/news/ai-is-destroying-the-university-and-learning-itself

Choice quotes:

"The result isn’t augmented intelligence but simulated learning: a paint-by-numbers approach to thought."

"Universities are being retrofitted as fulfillment centers of cognitive convenience. Students aren’t being taught to think more deeply but to prompt more effectively."

So many more spot-on observations. It's long. Actually read it, don't have AI summarize it for you.