Should be obvious but seems it’s not: don’t spam OSS maintainers or coworkers with AI code you’ve not reviewed yourself.
For coworkers only, sometimes fine explaining your testing and why reviewed isn’t necessary e.g. a one-time script.
https://simonwillison.net/guides/agentic-engineering-patterns/anti-patterns/
It’s hard to understate just how much more productive coding agents are at some tasks in YOLO mode. Essential to have a good sandbox for this, though. My favourite so far is sandvault: no Docker nonsense needed.
https://github.com/webcoyote/sandvault
The year is 2068. All matter in the solar system has been consumed for energy. All energy is used for the system’s sole remaining purpose: AI generation of another new frontend for Homebrew.
The author and I are convinced AI is net positive in engineering today but worth engaging seriously with the downsides.
https://tomwojcik.com/posts/2026-02-15/finding-the-right-amount-of-ai/
Google continues to take money for malware pretending to be Homebrew installation.
There’s nothing Homebrew can do about this. Google needs to fix it.
Please put me in contact with someone at Google high enough level to actually fix it.
https://github.com/Homebrew/install/issues/1074
Andrew nails here many parts of what actually makes OSS maintaining hard work.
Empathy is needed more for OSS sustainability than money.
https://nesbitt.io/2026/02/13/respectful-open-source.html
“This new technology will replace developers!” is not a new thing.
Nice look at what some previous claims were (and how they resulted in more developers and more software).
https://www.caimito.net/en/blog/2025/12/07/the-recurring-dream-of-replacing-developers.html
Great take about the cultural requirements to create “10x engineers”
https://randsinrepose.com/archives/sometimes-your-job-is-to-stay-the-hell-out-of-the-way/
31 January 2026
Lessons for non-Ruby projects on non-profits, governance, money and access in open source, drawn from the RubyGems dispute.31 January 2026
Homebrew 5.0.0 released in 2025. Walk through the major changes in 5.0.0, improving expectations based on other package managers and what they can learn from Homebrew's approach.
All the “faster Homebrew in Rust” projects are a bit like parsing HTML with regex.
The simplest use-cases seem to work, it’s easier and there’s just edge cases to fix.
Fixing these edge cases requires recreating Homebrew and using Ruby (which will be slower again).
27 January 2026
Interviewed by Screaming in the CloudMike McQuaid explains how Homebrew grew from a side project into macOS’s de facto package manager and how the project is sustained today.
This analysis was both helpful and hurtful.
Another reminder to focus on a single task and ship to completion whenever possible.
https://www.theengineeringmanager.com/growth/one-bottleneck-at-a-time/
16 January 2026
Interviewed by freeCodeCamp PodcastMike McQuaid joins Quincy Larson to discuss career lessons and the software engineering skills worth prioritizing next.
“The Failure Mode of Clever (is asshole)”
Applies to some OSS commenters I’ve seen…
Great take from John Scalzi (who also writes GREAT sci-fi books).
https://whatever.scalzi.com/2010/06/16/the-failure-state-of-clever/
13 January 2026
Like many people who now work with computers, I was told as a child I spent “too much time on screens” and then built a career out of it.
18 December 2025
I’ve been following what Justin Searls has been doing with his blog for some time. He’s been leaning into the “POSSE” (Publish on your Own Site, Syndicate Elsewhere) philosophy more and more. In practice, this looks like building your own version of a single-serving social network on your own site and exposing RSS/Atom feeds to other services to consume. Justin recently released POSSE Party which makes this easier by cross-posting to various social networks. I’ve complained for a while about (anti)social networking so I’m always up for new ways to use social networking less.
09 December 2025
The process of software estimation is frustrating for software engineers and those who consume their estimates. Consumers often ask “why can these software engineers not just tell me when it will be done?”.
24 October 2025
In tech, 3 years is often considered a “long tenure”. We maintain open-source projects for 2 years, then burn out. We start habits, lose momentum and quit.
21 October 2025
Interviewed by Software Engineering Daily
Mike McQuaid discusses Homebrew’s evolution, open source maintenance on macOS, and practical package management tradeoffs.
17 October 2025
Interviewed by Breaking Change - Hotfix podcastAlso available in swear-free/bleeped version on The Changelog and Friends podcast There will be bleeps.
09 October 2025
gem.coop was announced on Monday. As part of that announcement it was mentioned that I was helping gem.coop set up a governance process, continuing the work I’d first started helping with on RubyGems.
07 October 2025
Interviewed by GitHub PodcastHomebrew’s project lead Mike McQuaid joins Abby and Andrea to unpack what it really takes to sustain one of the most-used developer tools on macOS and Linux.
29 September 2025
Interviewed by Emanuel Maiberg on 404 Media
What happened to RubyGems, Bundler, and the Open Source drama that controls the internet infrastructure.
26 September 2025
When I first joined GitHub in 2013, there was no engineering management. They had people in engineering leadership roles (some with titles, some without) but no dedicated managers to check in with regularly. Initially I thought this was great. Over time, I realised it was actually pretty terrible. As a result, when I started my own company and was a manager for the first time, I wanted to ensure we provided “minimum viable engineering management”. What this meant was providing the necessary support and monitoring infrastructure to ensure great performance while letting managers, present (me) and future, spend most of their time on individual contributions.