Bookmarks: Week of Aug 31 - Sep 6, 2026

Blast Radio

shop.blastradio.com

This is crazy - relatively inexpensive dedicated internet radio broadcasting equipment. A wifi mic! and a thing to connect audio gear to internet radio. I’d never heard of Blast though, and golly their website smells of AI slop…

The platform for streaming live audio! Download the free Blast Radio app to start listening. Grab a Blast Mic, Blast Box, or Broadcast Software to start streaming yourself! All streams shared in high quality for clean, professional sound.

Tags: via:Hiro, internet radio, audio

DOC • We used to log off

doc.cc

Online used to be a place you could leave. We designed the exit out of it because an exit reads as a leak. Here is why building it back is a design responsibility, and what that means now that AI is dissolving the last seam.

Selected text

Now, online is not a place you go to. It is a layer you live inside. It clings to the morning, the dinner table, the checkout line, the bed before sleep and after waking. There is no visit because there is no return. What died slowly, with no announcement, is a concept we once took for granted: the session. A unit of time with a beginning and an end. Something that started, ran, and finished.

Tags: internet, via:Grant Potter, culture, online

AI Is Making Product Development Faster. But Where Did the Work Go?

medium.com

That time gives me room to stay with the problem longer. I can explore another direction instead of settling on the first reasonable answer, look for gaps in what I learned…

and

We already know AI can make certain kinds of work faster. What I find more useful now is understanding what kind of speed we are creating and what happened to the effort that used to sit around that work. Sometimes the work disappears. Sometimes it shrinks, changes shape, moves to someone else, or simply shows up later.

Tags: via:Grant Potter, AI, vibecoding, prototypes

AI Slop, with no Warranty - BCcampus

bccampus.ca

Online learning platforms in higher ed are now mostly enterprise applications on campus-scale licenses, leaving little room for small-scale experim…

Tags: BCCampus, person:Brian Lamb, vibecoding, webinar

GitHub - anildash/better-documents: A SKILL.md for Claude that applies communication best practices to AI-generated and AI-reviewed business documents · GitHub

github.com

A SKILL.md for Claude that applies communication best practices to AI-generated and AI-reviewed business documents - anildash/better-documents

Tags: person:Anil Dash, Claude, AI, documents, skill

Data Cookbook - Data Cookbook - Home

ucalgary.datacookbook.com

OIA’s tool to find data definitions, specifications, quality, systems, organizations, and community tools.

Tags: OIA, data, glossary, UCalgary

FBI Probes Service Selling 153M+ Drivers Licenses – Krebs on Security

krebsonsecurity.com

Thankfully, services like the ID verification thing that got hacked are just going to become more common thanks to age verification requirements etc. I think the boy may be in this because he had to use a service to verify his ID to unlock his LinkedIn account while looking for work…

Tags: person:Brian Krebs, privacy, identification

Comics in Education Minicomic as Syllabus! – Spin Weave and Cut

spinweaveandcut.com

Nick Sousanis (author of Unflattening, who did a postdoc at UCalgary a few years ago) just shared his course syllabus as mini comic, for a “Comics in Education” course he’s teaching. Wow!

Tags: syllabus, course outline, comic, person:Nick Sousanis

Agency and Agents - by Ethan Mollick - One Useful Thing

oneusefulthing.org

From the Hugging Face Incident to Twilight Factories

this overview of a recent OpenAI/ChatGPT incident where agents “hacked” the Hugging Face AI software repository is BA-NA-NAS. The agents demonstrated coordination and what sure sounds like altruism and self-sacrifice to benefit the collective. It’s all still sparkly autocomplete, but something weird is happening when it hits a certain scale?

The Hugging Face Incident demonstrates that autonomous AI agents can self-organize, coordinate, and perform complex tasks, yet they lack inherent safety mechanisms to seek human guidance. To ensure ethical and effective workflows, organizations should move away from fully automated “dark factories” toward Twilight Factories where AI agents proactively engage human expertise, judgment, and creativity.

Tags: person:Ethan Mollick, via:Scott Leslie, AI, Hugging Face, cybersecurity, OpenAI, ChatGPT, Kobayashi Maru