Cognitive surrender is starting to eat Amazon's "writing to think" culture of sense-making. Steve Huynh, via Substack (content warning : Substack platform).
https://alifeengineered.substack.com/p/on-the-death-of-amazons-writing-culture
Back in the 90s Seinfeld did a whole skit about bullshit circus stunt vapourware, only they called it “levels.”
Oh yes, Tesla *could* have shipped a Roadster at any time in the last decade. They just didn’t want to. And they could ship a so-called flying car in 2028. But they won’t.
A society is strong when its elders plant trees, knowing they will not live to enjoy the shade.
A society is weak when its elders eat its seed corn, knowing that they will not live to see the coming famine.
The first genuinely honest ad for generative AI I've ever seen, from Microsoft / CoPilot
A mediocre white guy sees his non-white colleague getting praised for good work, he asks AI to *literally copy it*, and then he gets all the praise instead:
www.ispot.tv/ad/SoDi/micr...
What's the OEIS number for the integer sequence { 4, 56, 66, 115, 155, 156, 230, 237 }?
“My name is Raganwald, hacker of hackers: Look on my Newton works, ye Mighty, and despair!”
Nothing beside remains. Round the decay of that colossal wreck, boundless and bare, the lone and level sands stretch far away.
"I was told," said the princess, "that the point of fairy tales is not to tell children dragons exist, but that they can be killed."
"Oh, absolutely!" said the dragon.
"You agree?"
"These rich, selfish, powerful, cruel beings that need killing, did you think they're literally dragons, princess?"
LLMs optimize for verisimilitude, not veracity:
“The model learned what confidence scores look like in text. It did not learn to make its scores match its accuracy.”
https://rumple.github.io/blog/2026/10/20/asking-llm-for-confidence.html
“State Corruption via Bailout” has being going on in Western Economies for my entire lifetime. Private profits, public losses.
——
p.s. I get that people are hurting, but black humour is the fifth stage of handling grief. Sending HugOps!
AI written text feels—to me—like Jazz without any dissonance.
Jerry Coker explained why I find that unsatisfying as a human: Coker’s theory was that as we listen to music, we can’t help but try to play along in our head. We get the most satisfaction when there is a balance between our finding the music familiar enough to correctly play along, and surprising enough to have us experience the joy of something *else* working.
Same with words for me.
Alan: "The only numbers that matter in a software design are zero, one, and infinity."
Georg Cantor: "א—?"
Alan: "Don't start with me, Georg!"
https://annas-archive.gd/md5/bc9f99b65b0dc4cd35b2ace8a02f829e
No regrets, Coyote.
No, really, these PHP vulns are quite serious and you should drop everything and patch. https://www.php.net/index.php
Charlie is taking my partner along the Great Lakes Waterfront Trail to one of his favourite beaches. Brilliant sunshine!
At some point in the 80s I encountered the concept that the only three numbers that should be applied to constraints are "zero, one, and infinity."
The argument being, for example, that a program that could open up to four windows was a poor design: It should open as many as hardware and operating systems tolerate. Or one. Or not have windows at all.
I don't know if this is a very real, durable principle that applies everywhere, but it has been a very pleasant heuristic.
Hugops to everyone running PHP
In 1987, Stewart Brand wrote “The Media Lab:”
https://annas-archive.gd/md5/fd36a2ea8f1d2006ff37083926328f5a
It predicted that media would become hyper-personalized via either:
1. Broadcatch, where user tooling aggregates and filters content to generate a medium personalized for every user according to their choices. This is #rss and the #fediverse.
Or 2: Narrowcasting, where a centralized service uses what it knows about users to generate a medium personalized according to the service's choices. This is algorithmic social media.
