2026-05 - W4

Being oncall taught me everything - Yao Yue

Without exageration, being oncall made me the engineer I am. … It taught me that being fast most of the time is not as important as being predictable, and that tail latencies are often far more important than median or mean in a system of scale. It taught me the value of simple and clear architecture in moments of crisis, the tradeoffs of common design patterns that textbooks never told you about, the interaction between an application and its environment, and the deeply connected nature of services in a datacenter. It taught me that operational excellence—thorough observability, consistent configuration, automation readiness, sensible default values—comes from good design decisions instead of last-minute patching two weeks before launch. My taste as a software engineer is that of a highly skeptical operator, and it has remained so long after I stopped being oncall.

A comma and a question mark

Interesting way of using local models to get bash commands.

And another blog post with code snippets: A Comma and a Question Mark Redux.

You should not update your dependencies in 2026 | Mendral

Nowadays, CVE and co are becoming more and more common… And the idea of “always be updating” is no longer relevant, and can even be harmful. Not sure what are the right solutions, if we no longer trust providers/dependencies.

A Love Letter to Neovim | Caio Bianchi

🫶 

Le DDD à l’épreuve de l’expérience utilisateur - OCTO Talks !

The Last Technical Interview. Today we will pour one out for the… | by Steve Yegge

Interviewing in the age of AI | dein.fr

Addy Osmani on cognitive surrender

Cognitive surrender is when you stop thinking altogether and blindly accept the answer what the AI gives you.

Get to the core of the thing

In every product conversation, the framing decides the discussion. People rise to whatever level of abstraction the question opens up. Board conversations especially. But it’s the same with executive staff conversations and team conversations. These framings are universally seductive because they make everyone in the room feel clever and sound smart, without requiring deep knowledge of your customers or creativity. You get the social status and the brownie points while avoiding the actually hard thinking of what are your specific bets on specific features. … The real question is not “wide or deep.” The real question is: what is going to work? What feature is going to resonate? What capability will get this person, the one whose call you took yesterday, whose frustration you can still hear, to actually buy, and to actually stay with your product?

We should be more tired than the model • Buttondown

All of these negate the supposed speed up effects of LLM-generated code in the short-term by adding friction, and yet, in the longer term, make me better at using the tool, because they solidify my own foundation instead of the foundation models’.

We should be more tired than the model.

And also another article (but this one, there’s a product to sell at the end though): I Didn’t Become a Developer to Review AI Slop.

Migrating from Go to Rust | corrode Rust Consulting

Really thorough article that compares the 2. It’s an honest take and really good writing.