About LockerGnome
Before LockerGnome there was Chris Pirillo’s Multimedia Madness, a home page where I promised at least one cool MIDI file every week. Finding the fun stuff was easy. Updating the HTML every seven days was a chore, so I let it go, and late in 1996 I started an email newsletter instead, which felt like writing one note to a whole bunch of friends I hadn’t met yet. Every issue carried a piece of software worth your dial-up minutes, and the people who read it were the ones who poked around in the registry because it was there.
Those readers grew up, and most of them never stopped tinkering. Some of you run the systems now (sysadmins, SREs, security folks, developers, the CTO who still has root on something), and the person who signs off on the monitoring contract on Friday is just as likely to spend Saturday stuffing a homelab into a dead router. You’re also the one everybody calls when their own tech breaks, whether it’s which phone to buy, why the printer quit or if that email from “Microsoft Support” is real. (It isn’t.) That’s why a cloud outage can sit on the same page as a new phone, a fix for the Wi-Fi that keeps dropping, or the computer you had in grade school running again in a browser tab.
When Apple ships a phone, Apple tells you all about it, and so do a few hundred sites reading from the same announcement. I don’t see much point in being the few-hundred-and-first. What you probably want to know is whether it’s worth trading in the one in your pocket, how it lines up against the Pixel and the Galaxy in a chart you can actually read, and what quietly went missing or got more expensive, which is rarely the part anybody puts on a keynote slide.
That’s also how I decide what to leave out. If the maker’s own page tells it just as well, there’s no reason for mine, which rules out most version bumps and anything that reads like the press release it came from. An outage waits until somebody can say what actually broke, because “a cloud region went down” happens most weeks and gives you nothing to work with. A breach needs a scene, a policy fight runs once a real admin is the one holding the install media, crowdfunding pitches wait until they ship, and a new app has to be more than another clone of an app you already have. When a story turns up on Reddit or Hacker News, I go find the person who wrote it up in the first place and link them instead of the thread that passed it along. And the headline tells you what happened, with no “you won’t believe” and no question you have to click to get answered.
How stories get made
Every story starts with something somebody published: a write-up by the person who was there, a project page, a court filing, a spec sheet. Each source is linked where the story uses it, so you can check my work. A TL;DR up top tells you in a few seconds whether to keep reading, and Steal this at the bottom hands you a couple of things to do about it, whether that’s a setting to change, a tool to try or a question to ask your vendor.
I use AI to help, and I’d rather you hear that from me. A language model writes the first draft from those sources, every claim in it gets checked against them, and anything that can’t be traced back gets cut. Then I pick what runs, edit it and put my name on it, and nothing goes live until I press the button. Quotes run word for word, credited to whoever said them and the outlet that reported them, and when I get something wrong, the fix goes on the story right under the headline (the corrections policy has the details).
The newsletter grew into a site and a community before LockerGnome went quiet in 2014. Now that it’s back, I can run a lot more stories a day than I ever could have typed in 1998, which makes deciding what’s worth your time the whole job.
Tips, fixes and arguments go to chris@pirillo.com.