Weeknotes: 2025-W31

Published: 4 August, 2025, updated: 4 August, 2025

Summary

A fairly productive week, but also a bit of a bland one staying at home pretty much the entire week.

Started working (again) on an AI texture synthesis proof-of-concept project. I actually built one last year, but it was a bit of a dead-end. Tried out using Python, trimesh, and moderngl for this, but found it wasn’t super-ergonomic in avoiding cross-thread OpenGL use in callbacks leading to a lot of time debugging. Restarted in Bevy and Rust, and sped through it without any cross-thread issues.

I finally got around to installing Tailscale, and I wish I’d set it up much earlier. It’s super easy, and now I can access a bunch of applications I run in VMs at home easily on any device from anywhere. As a bonus, it removes cert management pain and I can upgrade some HTTP stuff to HTTPS as well.

Reached some Beat Saber milestones this week - like my first ‘S’-tier Expert level rating, my first Expert+ success, and also managing to complete Sandstorm.

Photos

(Based more on time of processing than time of taking…)

Enjoying

• Reading:
  • Nightshade (Catalina #1) by Michael Connelly (★★★★☆)
    • The protagonist is a bit too reckless and operates with unearned antagonism against the system they’re a part of. Bosch and Ballard are also reckless and act outside the system, but it’s a lot more obvious that they’re doing that out of mission and necessity. Still, I’ll read the next one…
  • Wyrd Sisters (Discworld #6) by Terry Pratchett
• Watching (📺🏃🐌📅🎥📽️🏠🔄🏁★☆):
  • 📺🐌💑 The Librarians: The Next Chapter season 1
  • 📺🐌💑 Good Cop/Bad Cop season 1
  • 📺🏃 Sullivan’s Crossing season 2 (★★★☆☆)
  • 📺🏃 The Lazarus Project seasons 1 and 2 (★★★★☆)
• Playing:
  • Clair Obscur: Expedition 33
  • Beat Saber
  • Marvel Strike Force

Newly discovered

Helper-Scripts.com - Proxmox VE Helper-Scripts

While I’m a little skeptical of running scripts as root on my Proxmox server, the contents of these scripts as well as the coverage of different homelab software has already been useful to me. If you’re more willing to run this (maybe in a Proxmox-in-Proxmox VM?), I imagine it’s even more useful!

Highlights

My advice, for the moment:

1. Tune out both the most heated and the most dismissive rhetoric.
2. Focus on tangible changes in areas that you care about that really do seem connected to AI—read widely and ask people you trust about what they’re seeing.
3. Beyond that, however, follow AI news with a large grain of salt. All of this is too new for anyone to really understand what they’re saying.

AI is important. But we don’t yet fully know why.

– “No One Knows Anything About AI” by Cal Newport

This is probably good advice for anything new.

As I progressed through future projects, I learned to devote more time to thinking through the specifics of my ideas. The robots are good at guessing what I mean, but the less room I give them to guess, the less they need to dance. […] Rather than prototyping with code, I prototype in a spec. I explain to the robot what I want to build as a markdown file. This spec is the only thing we create. The process is no different than the first twenty prompts, except that the output is easy to read and easy to change markdown. The robots do a dutiful job of capturing my thoughts and their implications. No code. No APIs. Just writing.

And when I like what I’m reading, I ask the robot to build it. […] As a leader, I am surprised to find that improving my core skills in communication, setting expectations, and planning benefits both the robots and the humans.

Learning how to get the robots to dance for you will make you a better leader of both robots and humans.

– “Every Single Human. Like. Always.” by Michael Lopp

Like seemingly everyone, I’ve recently been trying out new “agentic” workflows a lot - almost exclusively, in fact - and talking to people who have been doing the same. There’s a strong correlation to those with stronger communication skills and mentoring experience and their reports of success. The people who report that agents can’t take on problems are often those who do best building things themselves - they rarely need to externalize their thinking to others.

Recommended

The Terribly Tragic, Totally Avoidable, Absolute Collapse Of The Gaming Industry - Acerola

A pretty good explanation of how the “AAA” and other investment-driven parts of the games industry have ended up following the same formulae and ending up with both big flops and great successes on games that don’t necessarily have more people playing or aren’t particularly innovative or even enjoyable to most people.

Hopefully we’ll continue to have great games being built by companies that aren’t chasing the latest revenue-chasing-at-all-costs trend, and naively I hope that more people will see through the companies that do chase those trends.

How Cutscenes Rekindled My Motivation - Night Run Studio

I appreciate how much can be done with appropriate choices and simple approaches. The urge to make things more is so strong, despite the biggest improvement is often creative use of a more constrained system.

I Built a Tool Post Milling Spindle - Marius Hornberger

Satisfying, but^H^H^Hand long.