Technology Commentary

Last week, Elizabeth Lopatto published an insightful article in The Verge. It boasted an intriguing title: ​“Silicon Valley has forgotten what normal people want.”​ “Within ... Read more

Genies give you code that’s a degraded facsimile of the mediocre code it trained on.

How do you fund work people don't want to pay for?

People love your work because of its impact, not how it works.

How travel has become an act of collection

The dashboard is green. The system is fragile. Taylor 1911, Goodhart 1975, Strathern 1997, Jeffries 2019: we were warned by name and kept the stopwatch under a procession of new names.

No.400 — Extrapolated futures archive ⊗ The dissonance is expanding ⊗ South Korea to spur a renewables revolution ⊗ Eye contact with a humpback whale

A meditation

No.399 — The term “AGI” is almost useless at this point ⊗ Frugal AI ⊗ Design futures in infrastructure ⊗ The AI revolution in math has arrived ⊗ First Indigenous Group to ban data centers from its land ⊗ A macro array of colorful slime molds

Are you frustrated with how the internet works? Me too. Today I’m going to think out loud about why. I’ve been writing for decades about what I’ve been calling “conversation…

Have you noticed all the folks bragging about the cool new tools they’ve hacked up using AI? In the last month or so, I’ve read newsletters from half a dozen or so people detailing vibe…

Friendster was the first social network. It has a long history, but the website friendster.com went dead in 2015 and the company officially…

One of the most important pieces of AI commentary: "software brain" is important to understand if we want to get through this era with our humanity intact.

Software brain is changing the world, but most people still aren’t buying.

Posted on Wednesday 29 Apr 2026. 1,087 words, 19 links. By Matt Webb.

GitLab, 31 January 2017, 23:30 UTC. Three hundred gigabytes deleted in two seconds. Eighteen hours later, the team had discovered that all five backup mechanisms were broken in different ways. Tales from the Bare Metal, Episode 01: a forensic walk through the most-cited operational postmortem in software engineering, and the principle the unixoid tradition has held all along.

I feel terribly guilty when I visit a new city, post photos of my travels, only to have a friend say "Hey! Why didn't you let me know you were in my neck of the woods?" Similarly, if I bump into an old acquaintance at a conference, we both tend to say "If only I'd known you were here, we could have had dinner together last night!" I do enjoy the serendipity of events like FOSDEM - randomly…

Code review is good at inspecting intent. It's structurally blind to a specific class of bugs - the ones where valid-looking data exposes wrong assumptions across layers. Here's why, and what complements review for input boundaries.

I have seen a common remark from people who are not the biggest fans of Odin and it is usually the remark that Odin is "full of sugar" which only works for the "blessed types" and it cannot be replaced or implemented for user-level types. Firstly, I don't think this is necessarily an example of "sugar" since that term implies it is shortening a construct / feature / idea into something smaller, whilst a lot of the ideas in Odin that would be classed as "blessed syntax" by such people would not be classed "sugar"...

Here’s what we’ve done—and what we’re still doing—to improve our availability and reliability.

Why predictable CSS state resolution matters, and how Tasty approaches it with state maps and non-overlapping selectors.

The defense industry lost the ability to make weapons when crisis hit. The same pattern is eroding software engineering skills. The timelines are identical.

A colleague told me something recently that I keep thinking about.She said, unprompted, that she appreciated seeing both sides of my AI conversations. Not

Interesting & joyful things from the previous week

ChatGPT has been publicly available for over three years now, and generative AI is woven into the tools students use every day: web search, word processors,

Current LLM infrastructure artificially limits developer control and system reliability.

The following article originally appeared on 'Dan Shapiro's blog' and is being reposted here with the author's permission.Companies are now producing dark

Message brokers look simple when one team owns them. The problems change when Kafka or RabbitMQ becomes shared infrastructure.

Livecoding Music with Fennel and Renoise