javier

@bartlebooth.bsky.social

data processor. recovering mathematician. lazy anarchist. tech mercenary. parent. immigrant. child of the corn. unpronounceable. toronto rodent. seeking puzzlement. Cuidado con mi corillo que somos un montón. nie mój cyrk, nie moje małpy

It is actually relatively easy to understand how they work. There are some mysteries on top that have to do with scale and generalization but those I feel are more about the structural characteristics of a large corpus of human language and what is truly encoded there.

Yan Basque@yanbasque.com · last wk.

Not gonna QT, because I don't want to add to the pile on this person is currently experiencing (even if deserved to some extent). But that argument drives me crazy. Just because *you* don't understand how LLMs work doesn't mean nobody understands how they work. This is a fiction.

We're currently growing intelligent machines. Nobody really understands how they actually work internally. That includes Gebru.
They have become demonstrably smart regarding general knowledge and programming in the last year or so.
2:35 PM • Oct 3, 2026

I hold these two views simultaneously: 1. The combustion engine is useful technology 2. Car-centric cities are dysfunctional and, on balance, harmful to the people who live in them I hold essentially the same view about LLMs

I was reading someone who put into words a way to discern ai writing (and which also makes AI writing annoying): that it tries to end every paragraph with a mic drop. I had mentally categorized this as "writes like a teenager trying to show off" but this is much more specific.

In a world where programming languages do not matter as much we may end up with only a couple that survive. A high level one, probably JS, and a low level one, probably Rust. Everything else may go away.

this is why trying to see far into the future will drive you insane, all you can see is how far away the world you want still is. but the more straightforwardly oppressive world of the distant past is also far away! because small people did small things that added up over time!

My @thestar.com column: Brad Bradford and Chris Alexander talk about Toronto's budget like it's some mystery we need outside experts to solve. But the mayor's spending record is all public, and tells a simple story of priorities. No mystery. Just math. www.thestar.com/opinion/cont...

Matt Elliott: There’s no mystery behind Mayor Olivia Chow’s budget record — there’s just math

Lately, the leading candidates taking on Mayor Olivia Chow in the race for mayor have taken to referring to the city’s budget like it’s some great cosmic mystery. A riddle

thestar.com

I predict a title like Muse by Meta made a teenager commit suicide or Muse by Meta is generating a new family of eating disorders among 13 year olds in about a year.

OpenAI is trying to frame their incompetence as technical prowess. Their robots are not becoming sentient. They are just badly set up in systems that are not properly isolated.