Simon @spoltier@social.cleverlibre.org

L. L. Bean stands for Large Language Bean. @lowqualityfacts

Don’t mind me unintentionally swearing aloud in the corner of the coffeeshop, just indulging in the black art of building software, everything’s… fine 😅

#Brutalist #graffiti
#photography

"If a developer experience initiative attempts to provide developers with new and adaptive strategies, but the larger context then invalidates the behaviors by which a person can execute those strategies, psychological outcomes may be more negative than not intervening in the first place. "

Geneva appreciation toot

What I'm listening to today: "Edelleen ja edelleen", Sleepers Tomb

Quiet, insistent drone ambient track. You're asleep, your phone's alarm keeps pushing at the barrier from some other world trying to break through and drag you out, but it's not working. A piece built up slowly on a modular suitcase that splays the track's internal process open to view like something on a dissection table. Good mood. I think the name is Finnish for "On and On"

https://www.youtube.com/watch?v=XhSssL8JxWw

I need a couple of dozen small blobs of similarly-structured-but-not-identically-structured JSON that look like they came from a biology lab to use in a tutorial on manipulating JSON in SQLite and Postgres. If you have such, I'd be grateful for a ping: gvwilson@third-bit.com. (Yes, I know, it's a weird request, but you're a weird bunch of people and I love you for it.)

There is really good stats thinking you can do on this and there are many models for mapping this type of change, but you won't get this thinking from business analytics or mainstream data science unfortunately. You need to look to the sciences that have done causal inference in complex real world situations.

People often say "ugh stop overthinking it. Just set a target and measure a change."

All models are wrong etc, but on some topics putting "simple" over everything is fundamentally broken.

Finally this is a great example of the dangers of thinking all changes are simple and linear. There are MANY patterns. A spike in negative evaluations is a well documented characteristic of any time you learn more about the world so again, a salient point made by @mekkaokereke that we must be very careful to observe before we diagnose. Here's an adjacent example in schools: sometimes we have evidence for an intervention working because it *slows down an existing negative trajectory*

Have you heard about code smells? It’s a kind of language for discussing suspect design and technical debt. Learning to recognize and name particular code smells could help team discussions and give ideas for good refactorings. https://youtu.be/L-cN7NI-Fes

are you normal or do you recognise if something is base64 encoded json from the ey at the start of the string

I think THAT happens because MANY of the things I can do to avoid wasting engineers' time are things that standard productivity proxies like "lines of code merged" would suggest are wastes of MY OWN time.

It's things like:

- briefing everyone before the meeting so we don't lose an hour to "round robin what is everyone working on"
- Socializing how to do a thing so it doesn't get done 45 different ways
- STOP people from doing the thing that a loud, misinformed voice is telling people to do

@HeyChelseaTroy I'm interested in hearing more about short circuiting status round robins with an intro summary