Weeknotes 2024.01

Hope you had a good start to 2024!

During the holidays, I turned off more than I had anticipated. The last year certainly took its toll on me and it seemed like my body was telling me to take a proper break.

Instead of delving into all the awesome ways a compiler can break when porting it to a new platform, I found some new sources for reading material (yikes!) and got way too deep into typefaces (again!). I think digital publishing and digital typography are such a perfect storm of design and technology, it simply took me hostage. Maybe I should give in…

Speaking of things I didn’t do, there have been quite a few topics amassed in the writing queue:

• An OpenAI leadership kerfuffle commentary; might keep it brief and link to all the commentary that came out since.
• On Open Source. 2023 exposed a lot of problems for Open Source software (OSS). I try to dissect it for everybody with only surface-knowledge of OSS and hopefully show it’s a societal issue.
• Return to Office: a product management perspective.
• Lightweight Markup Formats: writing down my research on the topic.
• Principles, Thinking & Strategy. Let’s get philosophical, shall we?
• OK Colours: some research in the new colour spaces in CSS.

And that’s just the short list with significant notes. The full list, including vague ideas, has 32 entries so far. Maybe I should ask an AI to type it for me 🤣. At one a week, I have 2/3 of year covered.

The design has also changed a bit again. I guess I’m still trying to find my voice for lack of a better phrase. So be prepared for more design and content experiments, until we settle into a groove.

Changelog

• Website: refactored in a lot of learnings, removed all of TailwindCSS (I will explain myself soon) and replaced it with OpenProps and vanilla CSS (there you have it). Seriously, what you can do with CSS props is amazing.
• Website: also included a few design changes that just felt right and unified the experiments.
• Website: at the same time there is still a lot to do, for example with footnotes.

Linklog

This time with a short end-of-year (EOY) comment.

AI

• Simon Willison has summed up the year 2023 in AI and it’s probably the best you will see.
• New York Times is suing OpenAI and Microsoft for copyright infringement. As I said in my article on the crawler issue, even if you don’t block an AI, it still can’t do whatever it wants with it, it has to abide by the content license and copyright.

EOY comment: if you want to use generative AI commercially, be sure to know what it’s trained on. The safest, but by no means cheapest, way is to deploy your own model and restrict access to trusted users, otherwise you have to deal with a multitude of novel security concerns. Beware of the AI bros (see Blockchain below).

Blockchain / Cryptocurrencies / Web3

• 2023 was the year of the fallen crypto bro, according to Ars Technica. I think if we could measure the bro saturation level of a technology, it would be a good indicator of what to be critical of.

EOY comment: most blockchain related activity, esp. related to cryptocurrencies, is all but dead outside of the core networks and it’s probably a good thing in terms of cleaning up the house. If you look at announcements in 2021/2022 by companies riding the crypto wave, you notice that the same companies fell silent or jumped on the AI band-waggon, in an ironic twist of fate forgot about their blockchain activities 😂.

2024 should be used to professionalise the narrative and purge the snake-oil utopia nonsense that our society is not ready for. Focus on legit blockchain uses, admit that, at the moment at least, decentralised regulation for cryptocurrencies doesn’t really work and detach the decentralised Web3 narrative from blockchain, because it’s only one way to achieve it (ever heard of federation protocols?).

Programming

• The New Stack had a Christmas present of a special kind: it was the first time I heard about the Christmas lecture of Donald Knuth, let alone that it was an annual thing. It’s laced with links to video playlists on everything Donald Knuth. Yes, including TeX. Who needs Disney+?!
• LinkedIn released it’s framework for developer productivity and happiness. I haven’t gotten beyond the overview, because I will not manage any engineers anytime soon. The connection between developer’s happiness and their productivity is old knowledge, but I’m happy it is still explored.
• Steve Klabnik dives into why memory safety is a red herring and why it’s about defined vs. undefined behaviour. If you’re interested in C++ and the self-proclaimed successor languages, this is a must read.
• Towards Oberon+ Concurrency. Some algorithm somewhere knows me well and my appreciation for Niklaus Wirth’s body of work, impact and design philosophies. How else could this have found me 😂.
• Shortly after reading Klabnik’s article above, which mentions Austral, the algorithm surfaced this recent interview with the creator of Austral. What caught my eye is that supply chain attacks are now affecting programming language design as well.

EOY comment: As I mentioned in Why Swift, we’re spoiled for options, even if they are more academic. The combined advances of how silicon is designed and increasingly complex and new applications leaves the field of programming languages and compiler infrastructures very exciting.

Policy

• I mentioned the EU Cyber Resilience Act and its implications for open source before. It’s still on my agenda to dive deeper into it. So it’s good to read from someone, who was actually involved to get a CRA that’s good for open source and their interpretation of it.

EOY comment: as with all legislation (or lack thereof), you have to wait and see until someone puts it to the test, even if it is sound in theory.

Automotive

• Xiaomi wants to be an automaker challenging the sorts of Porsche and much like their smartphone efforts, they let inspiration run wild. On first glance, you see design elements from McLaren, Porsche, Ferrari, Maserati and Acura/Honda. At least it’s not a design rip-off from one maker like with their smartphones 😂.

EOY comment: Xiaomi’s design aside, the relentless progress of Chinese car makers is astounding. Korean car makers are better at building cars for the European market than European brands, especially the German brands are shambolic, with few exceptions. The really good European brands sit on Chinese technology. BMW’s current design has been referred to either a beaver or sporting a famous-for-all-the-wrong-reasons moustache, depending on the colour of the grille 🤦.