Technology Commentary

When Anthropic introduced Agent Skills, I was initially a bit skeptical of the problem they solved–can we just use prompts and tools?–but I’ve subsequently come to appreciate them, and have explicitly implemented skills in our internal agent framework. This post talks about the problem skills solves, how the engineering team at Imprint implemented them, how well they’ve worked for us, and where we might work with them next. This is part of the Building an internal agent series.

One of the most useful initial extensions I made to our workflows was injecting associated images into the context window automatically, to improve the quality of responses to tickets and messages that relied heavily on screenshots. This was quick and made the workflows significantly more powerful. More recently, there are a number of workflows attempting to operate on large complex files like PDFs or DOCXs, and the naive approach of shoving them into the context window hasn’t worked particularly well. This post explains how we’ve adapted the principle of progressive disclosure to allow our internal agents to work with large files.

One of my provocative claims was that the lifespan of the chips may be significantly shorter than the accounting treatment given to them.

The wind and rain has swung east again and it’s stormy and very dark out, and so I’m out of the house and spending the day at the library which is such a nice bright and warm place to b…

Posted on Friday 26 Dec 2025. 430 words, 6 links. By Matt Webb.

The Architecture of a Tyranny

Ever wanted to just plug something in and conveniently read the hostname and IP addresses of a headless board like a Raspberry Pi? Chances are, a free USB port is more accessible than digging up a …

Recently, [Greg Kroah-Hartman] proclaimed the joyous news on the Linux Kernel Mailing List that stable General Purpose Interface Bus (GPIB) support has finally been merged into the 6.19 Linux kerne…

Although metal alloys is not among the most exciting topics for most people, the moment you add the word ‘radioactive’, it does tend to get their attention. So too with the once fairly …

Every Useful Thing I’ve Learned About Databases—and Now AI Coding Tools—Started with Failure.

Speeding up takes just one word.

Two articles hit my feed this week that got me thinking: Simon Willison's piece on delivering proven...

Prusa released a new set of Core One files on December 19th under a new license they made called Open Community License (OCL). Their blog post explains

To the surprise of almost nobody, the unprecedented build-out of datacenters and the equipping of them with servers for so-called ‘AI’ has led to a massive shortage of certain component…

I don’t connect with the current tech landscape.

A long time ago, there was a big difference between PC and console gaming. The former often came with headaches. You’d fight with drivers, struggle with crashes, and grow ever more frustrated…

Why my worst database schemas made me a better engineer — and how that translates to AI coding tools.

Replacing DevTools' HTTP Handler With Redis Pub/Sub

Working on a codebase or using an application itself can sometimes be challenging without proper documentation. One effective way to bridge this gap is by using end-to-end (E2E) tests as a form of documentation. E2E tests not only verify that the application works as expected but also provide concrete examples of how different features and functionalities are intended to be used.What is E2E Testing?End-to-end (E2E) testing is a software testing methodology that involves testing an application fr

Turning billions of everyday devices into the Internet's crowd storage layer. Decentralized, autonomous, and quantum secure. Strong, private, connected data. And you hold the key.

WordPress has the ability to empower 43% of the Internet with AI.

Last month, I made the following remark on LinkedIn about the incident that AWS experienced back in October. To Amazon’s credit, there was a deep dive talk on the incident at re:Invent! OK, i…

Al and I were talking on the podcast about the Home Assistant home automation hub software. In particular, about how devilishly well designed it is for extensibility. It’s designed to be added on t…

The Model Context Protocol (MCP) is genuinely useful. It gives people who develop AI tools a standardized way to call functions and access data from external

Technology should bring out the best in humanity, not the worst—a manifesto for resonant computing built on five principles that reject hyper-scale extraction for human flourishing.

We noted that Excel turned 40 this year. That makes it seem old, and today, if you say “spreadsheet,” there’s a good chance you are talking about an Excel spreadsheet, and if not,…

Before you break apart your monolith, understand the hidden costs, perverse incentives, and cultural traps of the microservice architecture.

Writing Code vs. Writing Prose Programming languages were developed because natural language isn’t adequate to communicate instructions to computers. Unlike natural language, programming languages have strong composability: you can take a line of code and start putting things into it or removing things from it to drastically change its meaning. Doing that with natural language is also possible, but not the extent that it is with a programming language. With code, operations can be stacked, complex behavior (which would take dozens of sentences to explain properly in natural language) can be represented with a couple of lines of code, and special notation/syntax is developed to aid readability.