Codex bundles LibreOffice I was poking around in my ~/.cache/ folder using OmniDiskSweeper when I spotted something interesting. The OpenAI Codex desktop app (since rebranded to just ChatGPT) has 1.7GB of stuff in …
Pi vs OMP (Oh My Pi Agent): Everything You Need to Know | Composio We benchmarked Pi vs Oh My Pi on 30 real-world tasks. See which coding agent won on performance, speed, tooling, token usage, and cost.
Why your agent is better than mine Why Railcode doesn't ship an in-product AI agent at all, and why offering the choice would make the product worse.
I Built an Agent Harness to See What's Inside I couldn't explain what the harness around a coding agent actually does, so I built one in Go. What goes into the loop, which problems the existing harnesses had already solved, and whether the transparency was worth it.
How accurate have Ed Zitron's AI skeptic predictions been? I was curious how well the predictions of the most widely cited AI skeptic I've seen (Ed Zitron) have done, so I looked at how his predictions panned out. To disclose my own biases, I've never had a particularly strong pro or anti AI progress position. Fo...
Building the Open Task Light Steven Bennett just launched the Open Task Light build kit after four years of publicly documenting his work on it. Today’s launch video includes a good summary of the project. In recent months, I’ve been helping beta test the kit and the build documentat...
Three kinds of contradiction manual review misses Formal requirements inspections average 78% defect removal, the weakest of any inspection type. Why contradictions survive review, and more reviewers don't help.
I Built My Agent Computer in the Cloud - Hadi Javeed's blog Notes about startups, AI, health-care and overall engineering
The failure your dashboard can't see The first time I pointed a load generator at a hosted inference API, I learned something about my own measurement before I learned anything about the provide...
The Browser's Main Thread Is Expensive | kciter.so What comes to mind when you hear "frontend optimization"? For most of us it's things like reducing network requests, shrinking the bundle,…
On not becoming a cyborg I was a voracious reader as a kid, and I noticed a funny phenomenon: whatever writer I had most recently read, my next paper for English class would sound like them. If I read a bunch of Stephen Ki…
Learnings from SecAgent Development Insights and lessons learned while building and optimizing SecAgent—a cybersecurity alert triage agent built with PydanticAI.
Adverse Selection & Markouts Being a market maker isn’t for the faint of heart. Every quote you place is at risk of being picked off if you aren’t careful. In this article, I cover the concept of adverse selection, as well as how markouts can be used to measure it.
Lambda Land | Two Ends of Programming Some years ago I worked with a software developer from Germany. I was an intern at the time. More than once, after I asked this engineer for a review of a pull-request, he’d tell me, “Hauptsache es funktioniert” (translated: as long as it works). I have h...
[Java][JVM][Tuning][Profiling][G1][JIT] Why is Arrays.fill 265 times slower on G1GC? [Java][JVM][Tuning][Profiling][G1][JIT] Why is Arrays.fill 265 times slower on G1GC?
How skills and MCPs actually work The AI ecosystem can be really confusing: Agents, MCPs, skills, plugins, everything is an overloaded term and a fuzzy concept. This is an explanation of how these constructs work in plain terms, useful if you're an engineer wanting to go from passively co...
The Endless Temptation of Claude - Discard Pile This post is inspired by the creator of Paint.NET announcing that he has added WINE/Linux support to his art program by using Claude to code it, leading users to have...
Aurora Store returns a “&$Server busy, please try again later.” error when trying to install an application (#1566) · Issues · Aurora OSS / AuroraStore · GitLab Description Currently, Aurora Store, including Nightly (2026-08-31), returns a “&$Server busy, please try again later.” error when attempting...
Introducing Claude Fable 5.1 and Claude Mythos 5.1 Our most advanced models for coding and knowledge work. Their research capabilities also offer an early glimpse of how AI models will contribute to scientific progress.
[138] Artificial Deadlines (Part 1): Evidence of Fraud in an Influential Study About Procrastination A new paper in Psychological Science (.htm) reports a failure to replicate Study 2 of Ariely and Wertenbroch’s influential article entitled, “Procrastination, Deadlines, and Performance: Self-Contr…
A Walkable ASCII Cyberpunk City in One HTML File Update on the project: https://www.youtube.com/watch?v=UCKEDWowc0o Try the prototype: https://ko-fi.com/s/e1e0f91951 I wanted to build a city that feels alive, but is made entirely from ASCII characters. This is a small custom engine built with JavaScr...
