12 October 2026

25 links · Monday

Yes I Understand How LLMs Work

I continue to hear this claim made regularly that “nobody knows how large language models work,” and I want to put this claim to rest finally.

On Intelligence and Sentience

I have a conjecture which I’m almost certain is true. Intelligence — as defined as ability to solve open-ended, novel problems — 1Yes it’s my em-dash if at all necessary, is not strictly the same as sentience, which — without defining it — consists of more tha…

Immigration

A border is where the control of one sovereign ends and another’s begins. In the ideal case where individuals are sovereign, borders are in the simplest case just the boundaries of your home.…

Muse Wonder

A frontier model that leads on personal and professional agents

Release: datasette-ports 0.4

Find all currently running Datasette instances and list their ports

GLM-Legal: Advancing Agentic Legal Intelligence

Large language models are shifting from single-turn code generation toward broader professional collaboration (Professional Cowork). Real-world knowledge work introduces distinct requirements: models must digest fragmented source files under concrete objective…

D-Bus Hell

The one where I finally address the number one comment.

DNSSEC Validation Failed | no dogma blog

Today I started getting DNSSEC validation failed errors while using Ubuntu 24.04. All DNS lookups were failing! Here is how I fixed it.

What's Wrong with GitHub - GNU Project

Software developers, the GNU Project urges you to avoid hosting your repositories on GitHub. This is to avoid being on the receiving end of harmful practices, and avoid leading others to be victims too. This article focuses on a few of GitHub's worst problems,…