Daniel

I don't celebrate the holidays by spending money. I celebrate the holidays by spending time. Community over consumerism.

"One step forward, two steps back."

You guys get steps forward?!

YouTube video by Hank Green Give me a single reason why Sora2 should exist.

Love when @hankgreen.bsky.social makes a video screaming what I have screaming in my head. It's good to not feel so alone.

I'm just eight instances of Claude Code in a trenchcoat

A t-shirt showing a pile of tires on fire, saying in large typeface: “fight fire with tire”, with a small byline below saying “SRE: fixing software with more software since January 1st 1970” There are also 3 round stickers: 1. A bus labelled “SRE” dumping shit off a bridge onto a boat (in reference to the Dave Matthew’s bus tour incident), framed with text saying “Shitposting Reliability Engineering” 2. A bee on fire framed with text saying “honeycomb on-call incident commander” 3. A computer on fire with multiple bees flying around framed with text saying “honeycomb on-call incident responder”

New SRE team swag is in

The Leftover principle is a consequence of “automate what you can and what is left over is done by humans”, implying people are left with trivial tasks not worth automating and tasks too rare or complex to automate.

Assuming this can be done without negative impact to leftover tasks is often wrong.

Interesting new feature in #Claude Code v1.0.86
/context 👀

Solopreneur playbook in 2025:

- Find a painful problem
- Sketch the solution
- Break it into logic
- Hand it to Claude Code
- Ship fast
- Iterate from real users

That's it.
No 50-page business plan needed.

#SaaS #buildinpublic

Trust

Day 1/30 sharing my favorite doodles and drawings.

I am planing to take the same journey, care to share a primer I can use in my attempt to spate some yak shaving?

Proposed talk idea: skill transfer through storytelling

Cognition is a team sport, with equipment

Claude Code usage graphed in Honeycomb show the P50 and P90 of token usage by session and by model.

You can just graph your Claude Code usage using OpenTelemetry on Honeycomb.

ok I had to do it, tag urselves

Teach a robot to fish and soon you won't remember how to.

Root cause: didn’t ask the LLM to think hard enough

It’s funny how LLMs are fully discrete in their input/output, and yet their internals are all floating point, which is a discrete approximation of continuous values. It’s digital on the outside and an approximation of analog on the inside.

‘We’re Definitely Going to Build a Bunker Before We Release AGI’ The true story behind the chaos at OpenAI.

In a preview of EMPIRE OF AI for @theatlantic.com, I share new behind-the-scenes details on what happened when OpenAI’s board fired Sam Altman—as well as the core argument of my book and why it matters to everyone. Gift link.
www.theatlantic.com/technology/a...

You're on a landscape, a certain height above sea-level. There's a slope, a gradient, to the hillside you're on. Left higher, right lower. Which way do you go?

Depends on where you're going, right?

(Unca Nat 🧵)

YouTube video by Code Sync Keynote: The Socio-technical Elements that Make Good Platforms - Charity Majors & Fred Hebert |CBA25

another banger by @charity.wtf

"When everyone is looking at a different view of reality, that's what creates silos."

There are, in principle, two different planning strategies. There is planning in order to maximize gain, and there is planning in order to minimize disaster.

- Ursula Franklin,
The Real World of Technology

Given the world around us, I have never been more grateful for the lesson drilled into me by ADHD coaching & practice:

🌼YOU own your attention.🌼
🌸no one else does.🌸
🌺no one else can MAKE you get distracted.🌺

Here are some of the tools I lean on heavily; I would love to hear yours.

Those are really good insights, thank you for cutting through the hype and finding practical patterns for everyday engineers to explore.

I can only wish engineer leadership at large would have this narative instead of buying into the smoke and mirrors.

The Four Patterns of AI native Development - review The Four Patterns of AI native Development Introduction The job of a developer consists of performing many different tasks. Like any new technology, AI helps us to execute these tasks easier, faster...

Preparing a blogpost on AI native Dev Patterns - Here' a preview - Feedback and comments are welcome !

docs.google.com/document/d/1...

Restructuring How We Think About Alerts Rather than focusing on the events that result in alerts, Fred proposes we look at the system and center it on the operators instead.

New work blog post, covering some re-framing about alerting.

Make a distinction between: alarm, anomaly detection, fault identification, and corrective action.

Consider the cost & necessity of validating each, that they can all storm you, and maybe then willingly reduce how many can disrupt you.

When you look at a boxes-and-arrows diagram, key in mind that nobody “owns” the arrows, only the boxes.

It's so embarrassing to read a book. You should already know the information that's in there

i want to be exactly rich enough to buy a spacious, perfectly organized kitchen, where i’m not forced to stack all my pots and pans in a psychologically unmanageable way. and no richer

At your company, is your key issue collaborating on things that could be transactional, or trying to treat collaborative things transactionally? Or somewhere in between?

This sounds simple and obvious, but it is *not* how most teams are used to debugging their systems.

My favorite analogy for instrumentation is like a headlamp. 🔦 The best way to improve your systems is by instrumenting in front of your face. Anything you're trying to understand, instrument first.