9 May 2026

28 links · Saturday

Please Make Mistakes

When your team adopts AI, mistakes are inevitable and necessary. Your job as code owner is to catch them before customers see them, and improve your instruction files so they don't happen twice.

Alex-Dahl

I have been a manager for quite some time. I have seen plenty of different behaviors play out. Eagerness. Apathy. Rage. I thought I could navigate all of them and make sure they could direct their feelings to align with the organization's goals. I saw myself a…

The Mirror Is Part of the Machine | Yusuf Aytas

Why telemetry explodes into cost and noise, and how to fix it: treat every signal as a decision, not exhaust, and govern ownership before it's too late.

I still like Jenkins

Before GitHub Actions, before GitLab CI/CD, and before the panoply of startups all promising to help you build your software faster and with less hassle, there was the humble, open source Jenkins. Or perhaps I should say is: Jenkins is still actively maintaine…

Money is renewable. Time isn't.

We built a whole industry to count our dollars. We barely count our hours. The asymmetry is backwards.

How we accidentally recreated old FB

Background We started off with a simple idea for a product; we wanted to re-use the chat interface for everything except chat. It was supposed to be just for photos but expandable later. The idea was that most people already share their photos with others on W…

Lies, damned lies, and Elastic's benchmarks

It's a spicy title, but one that is warranted. Elastic recently published a blog post titled 30x faster than Prometheus, that claimed that Elastic's new TSDS timeseries engine is much better than Prometheus. When I started reading it, I was curious; maybe ther…

Laptops all have built-in security tokens these days

I’ve been a fan of security tokens for a decade now and have accrued quite a collection. This redundancy isn’t a bad thing, as security tokens are easily misplaced and the only way to recover from a lost token is using a second token that is also registered wi…

LLMs Corrupt Your Documents When You Delegate

Large Language Models (LLMs) are poised to disrupt knowledge work, with the emergence of delegated work as a new interaction paradigm (e.g., vibe coding). Delegation requires trust - the expectation that the LLM will faithfully execute the task without introdu…

A recent experience with ChatGPT 5.5 Pro

We are all having to keep revising upwards our assessments of the mathematical capabilities of large language models. I have just made a fairly large revision as a result of ChatGPT 5.5 Pro, to whi…

AI is Breaking Two Vulnerability Cultures

A week ago the Copy Fail vulnerability came out, and Hyunwoo Kim immediately realized that the fixes were insufficient, sharing a patch the same day. In doing this he followed standard procedure for Linux, especially within networking: share the security impa…

Cartoon Network Flash Games

Explore a collection of Cartoon Network browser games inspired by The Powerpuff Girls, Dexter’s Laboratory, Samurai Jack, and other fan-favorite series.