22 November 2025

23 links · Saturday

Tom Forsyth (@TomF@mastodon.gamedev.place)

Attached: 1 image Recent discussion about the perils of doors in gamedev reminded me of a bug caused by a door in a game you may have heard of called "Half Life 2". Are you sitting comfortably? Then I shall begin.

Agent Design Is Still Hard

My Agent abstractions keep breaking somewhere I don’t expect.

From Bottleneck to Multiplier

TL;DR Engineering leaders can multiply team intelligence instead of bottlenecking it. This series introduces a framework with five interconnected pillars: psychological safety, clear success criteria, intentional delegation, alignment through documentation, an…

Your Codebase Is Probably Fighting Claude (part 1) - ambient-code.ai

I have been trying to keep our genAI workloads “in the pocket” of a maximum effective context window.  For performance, accuracy and the second order cost control benefits from those. Things work beautifully, but it’s not at all efficient (wrote about an effic…

Is Matrix Multiplication Ugly?

A few weeks ago I was minding my own business, peacefully reading a well-written and informative article about artificial intelligence, when I was ambushed by a passage in the article that aroused …

Why I wrote Rmlx

I’ll start on a downbeat note: clearly, if there ever was a contest between R and Python for dominance of machine learning, R has lost. Nobody is writing AI in R.

Home Row

112 WPM. Take that, typing teacher!

Round Two

I'm excited to share that I am officially a founder again! This time, I'm building something close to heart: not a healthcare SaaS platform, but a tool th...

50 Things I Learned in the Last 2 Years

1. Be very specific when returning data from the database. Don’t blindly select all fields. Only return what’s needed. 2. Many people get way further ahead in life than you with fewer skills simply because they have the courage to try. 3. Communication is…

Hyperscalers are hard

How NFTables Drives Kubernetes Nodes to Memory Exhaustion at 100k Scale OR How Kubernetes optimizations become catastrophic at hyperscale.