Weeknotes 2024 W37: Finding problems
Quick bits:
-
The weather is finally much cooler! I am glad to not have to endure 31°C (89°F) heat anymore.
-
Last weekend I started a new acting class, but I think it’s too beginner-focused. I might not get much out of this one, unfortunately. We’ll see.
-
The acting class has some scenes that I find to be not very well written. That made me wonder: could I write some scenes for acting classes? That’d be fun.
-
I don’t understand what the hell “AI” is for anymore. We’ve built machines that are so complex that they are beyond our understanding, are utterly unreliable, and are extremely inefficient and power-hungry. Someone explain to me what the value proposition is, exactly?
-
I spotted my first Grünpfeil1 in the wild, allowing a turn right on red (but in this case only for cyclists).
I am struggling with the one-person play that I am writing.
The format I had in mind until now would’ve been difficult to pull off, at least as a one-person show. I didn’t want it to be a giant monologue (that’d be boring), but playing different characters simultaneously is not something I can confidently pull off.
So I’ve pivoted, and the idea I’m trying out now is a handful of vignettes of people presenting their own viewpoints and experiences of the events in the story, rather than have the story itself play out on the stage.
This feels considerably easier to write, and it has the benefit of not needing multiple characters on stage at one time. It also creates more of a sense of mystery, because the audience is now a (small) step further away from the real story that underlies what is being told on stage.
But it’s not all roses. How do I tell a story through a series of vignettes that is engaging, does not repeat information, and keeps increasing the stakes and tension? Doing so is essential to keep the audience engaged.
Also: I don’t think my acting skills are even close to being able to pull off playing multiple distinct characters. That’s not something I’m worrying about yet — there is no time pressure so I can take all the time to make this work.
This week, I’ve come to realize that I am rather good at finding problems in my day job as a software developer. I like to think that this is because I am thorough in the work that I do. A few examples:
-
I did an internship where my task was to make an existing abuse-detection system for a social networking site2 production ready. I ended up, instead, proving that the entire system is unusably broken3 and that doing any work on it would be pointless.
-
I once found out by accident that the QA team had been testing the wrong built artifacts. Due to an overly complex build pipeline, the artifacts being tested were not the ones being released.
-
In the course of my daily work, I have found data breaches on multiple occasions, and (of course) escalated them.
-
One time, a tiny new-joiner task ballooned into a month’s worth of work of bug fixes and refactorings.
I enjoy diving deep into topics. I want to fully understand the topic I am working on, and create solutions that are complete and correct. I live by the phrase “if it’s worth doing, it’s worth doing well.”
But I have on multiple occasions been told off — and even shouted at — for exposing problems. I believe that finding and flagging issues is the right thing to do, and certainly not something to respond to with outright hostility.
My best guess is that raising problems is not appreciated because creates unscheduled work — you can’t ignore the problem after it’s been escalated, and ignorance is comfortable.
Entertainment:
-
Jim Jarmusch’s Coffee and cigarettes4 is a fun collection of vignettes. The quality’s a bit mixed; some of the stories fall flat, but others are excellent — Cousins? with Alfred Molina and Steve Coogan is perfect.
-
With Satisfactory5 now having reached 1.0, I picked it up once again, in a brand new save.
It is certainly an addictive game. But especially with the 1.0 release, I’m impressed by how balanced the gameplay is. There is plenty new in this latest version, but every single nice new thing comes with a cost that is just right.
Links:
-
Are Taiwan’s Roads Still a “Living Hell”? (Not Just Bikes): Oof — I complain about the traffic in Berlin often, but Taiwan is/was orders of magnitude worse.
-
‘Celebrity Number Six’ Internet Mystery Is Solved (Jason Koebler for 404 Media)
-
The weirdest water bottle you can buy (Steve Mould): I want one! For no particular reason.
-
How To Write a Good Story That Doesn’t Give Me Headaches (Jeff Vogel)
-
How to Monetize a Blog (modem.io): That blog post… definitely wasn’t what I was expecting.
-
AI and the American Smile (Jenka Gurfinkel)
-
I wrote about this in Weeknotes 2024 W30: Traumatic response. ↩︎
-
The system’s machine learning model was tested on its own training data — a rookie mistake. ↩︎
-
Coffee and Cigarettes, written and directed by Jim Jarmusch (Asmik Ace Entertainment, BIM Distribuzione, Smokescreen Inc., 2004). ↩︎
-
Satisfactory (Coffee Stain Studios, 2024), published by Coffee Stain Publishing. ↩︎