Weeknotes 2024.07

We end a week of two days off and otherwise being very busy and productive. Preparations for the company went along nicely. First workshop in a while is coming this week, so I have been busy preparing boards left, right and center.

Squeezed in a few minutes with Zed. The collaboration features are the most interesting things. The license, while open source, is a restriction in my view.

Changelog

• ✨ New & Added
  • Article about BackstageWorks
• 🛠 Improvements
  • Website: a few design tweaks to make the site more consistent and easier on the eyes
• 🔬 Research
  • Adding D3.js to my portfolio proper
  • Variable fonts & tools in open source. Some call it research, I call it a rabbit hole 🤣
  • Anything new in product discovery? Short answer: no.

I still haven’t managed to fix a bug in Chrome based browsers that eludes me 🤷‍♂️

Elsewhere

The industry is having some important rethinks. When introducing something, always think of internal vs. external complexities. Internal complexities are inherent to the problem you’re solving, external complexities are added complexities due to the way how you solved the problem. Minimise the latter. Unfortunately, a lot of alleged “best practices” are falling into the “added complexity bucket” and are unnecessary in some (or most) circumstances.

We are at the point of the pendulum, where you may see more and more ‘backlash’, which is code for challenging a status-quo.

Rethinking X at The New Stack

The New Stack dropped three articles in short space that have something in common: Rethinking Observability, Rethinking Testing in Production^[1], and We Need to Rethink Risk in Vulnerability Management. Whether you agree with the articles or not, I like the framing much better than the “X considered harmful” we had a decade ago.

Stopping to rethink and reassess is a good thing. Expect that your favourite tool might take a hit, listen to the feedback and decide whether a course correction might indeed be necessary (before external factors force you to).

jQuery 4.0.0 BETA

I want to add to the sentiment that the developers behind this release are legends! They have a job to keep a big part of the web going, which is still using jQuery, while the dev community has largely moved on to greener pastures. Since we see quite a bit of a simplification movement in the web dev space, there may be a renaissance looming for jQuery.

Anyways, the release post is pure gold if you like dirty little details. The best part is about the focus event orders:

Starting with jQuery 4.0, we no longer override native behavior. This means that all browsers except IE will follow the current W3C specification, […].
For those that are curious, the W3C specification previously defined a different order […] [but] few thought that intuitive and the spec was changed in 2023 to match what browsers have already implemented. Ironically, the only browser to ever follow the old spec was Internet Explorer

(Emphasis mine) I guess you can’t win when you’re called Internet Explorer, even if you follow the standard…

Also interesting: while jQuery may be considered an “old-school” way of doing things, their tooling to get the best jQuery possible out, is anything but. I have a whole new appreciation for the project. Kudos!

Rust-C++ Interoperability is worth $1 Million to Google

With both Microsoft and Google driving interoperability for Rust, this should make progress leaps and bounds. They can even take inspiration from Swift, which already made some progress in this field.

Apple releases iTunes replacements for Windows

It’s amazing to note that iTunes for Windows outlived its Mac counterpart by four and a half years. Now it’s time for new iCloud, Apple Music, Apple TV and a brand new Devices apps. Yes, there are people who use iPhones/iPads and Windows PCs, get over it.

While that’s good news for Windows users with an iPhone, I will add “Cross platform SwiftUI includes Windows” to my WWDC bingo card this year.

Apple releases new configuration language

Speaking of bingo cards, I have to agree with Daniel Jalkut, probably no one had “Apple will release a new language, implemented in Kotlin, with IDE integration for everything but Xcode” on their card.

Pkl is a programming language for configuration. The memes are plentiful, including XKCDs take on the 15th standard, however I also see some positive takes on it. (via)

Why Kotlin? Apple has plenty of Java running on their backends, so it makes sense to modernise it via Kotlin. Btw, Apple also uses Rust. You see, different technologies can co-exist.

Mozilla switches CEO

Mitchell Baker steps down as CEO, and Laura Chambers takes over in the interim. The numbers are bleak, except when you have been the CEO, but there is absolutely no difference in other companies, so get over it.

There is plenty of commercial activities Mozilla can do in order to get less reliant on Google’s money. They probably should get a move on, because I’m not sure how valuable that search box default will be in the long run.

AWS releases experimental JavaScript runtime

Low Latency Runtime is designed for Serverless and released as very permissive MIT-0. 👏

The AHA Stack

AHA stands for Astro, htmx and Alpine.js and strives to make frontend development simpler. It sounds much more counter-culture to the German ear, where “Aha!” can be short for “I told you so!” or “I knew it!”.

Full disclosure: I’m in the EHA camp, who use Eleventy instead of Astro, because it goes to 11.

Common Git Mistakes

Julia Evans compiled a list as a Gist. (via)

Infrastructure decisions

The title says it all: (Almost) Every infrastructure decision I endorse or regret after 4 years running infrastructure at a startup. I wish we could have more of this, so I’m aspiring to do the same in the future. (via)

Best take on the current AI capabilities

By someone who should know. I quote the same bit as my source:

Reality is that LLMs are not AGI -- they’re a big curve fit to a very large dataset. They work via memorization and interpolation. But that interpolative curve can be tremendously useful, if you want to automate a known task that’s a match for its training data distribution.
Memorization works, as long as you don’t need to adapt to novelty. You don’t need intelligence to achieve usefulness across a set of known, fixed scenarios.

Seven Trillion 😂

Sam Altman seeks $7 trillion in funding for AI chip fabrication. According to the same article, the entire semiconductor industry combined makes about $527 billion, or $0.5 trillion, in sales. This makes the number both realistic, considering they have to build thing up more or less from zero, and hilarious at the same time.

To put this into perspective, we’re talking major state-run initiative levels, considering the Apollo space program, which brought us to the moon no less, was $178 billion (in 2022 dollars).

Seven f—in’ trillion…

Is our society too complex to survive?

Big Think asks a valid question. For someone trained in modelling complex systems, I’m worried as well. I guess we will survive, but have reached the point where we have several issues:

• Over-specialisation is increasingly harming adaptability.
• We added so much complexity to the system we live in that it’s harder and harder to tweak any parameter, without getting unforeseen consequences, usually bad ones.
• To handle the complexity, models are simplified by ignoring hard to quantify parameters for various reasons.

The thing is, you observe this even on different levels, e.g. companies can serve as a subset of the society and boy-oh-boy, executives are not well equipped to handle this complexity 😱

Wanna feel old?

Star Wars I: The Phantom Menace is returning to theaters for its 25th anniversary. Including the favourite Star Wars character of all time, Jar Jar Binks!

That’s all. I skipped on quite a few other interesting bits and pieces. What do you think about this style of linklog? Let me know via LinkedIn or Mastodon. Starting next week, I will try yet another way to structure the Weeknotes a couple of times, before settling on one.