If you only read one thingThe magic of software; or, what makes a good engineer also makes a good engineering organization (14 minute read) A long and enjoyable read, best for Sunday afternoon. The main point seems to be, that we believe that software development is Vision → Engineering, we have an idea and then it is build, but that engineering is shaping the vision. The article ends with “The qualities that make a good engineer are often the same ones that create a good engineering organization. Both start with deep understanding as the basis for innovation – cultivating the curiosity to look inside the black boxes. Both recognize that vision and engineering are often deeply intertwined and mutually informative rather than linear. The magic of both software and software organizations comes from those moments where insight into how something works sparks entirely new ideas about what it could become.” Go read it. https://moxie.org/2024/09/23/a-good-engineer.html
Stories I’ve enjoyed this weekHow an AI-enabled software product development life cycle will fuel innovation (18 minute read) McKinsey, I know, I know! BUT an interesting read, it predicts the 4-step life cycle of software development with Discover, Validate, Build, Launch and Scale might become two-step: (1) Discover, validate and experiment and (2) Build, Launch and Scale. Might be, we will speed this up is sure, but McKinseys guess is as good as mine. Good read overall, what I disagree with, “The idea of PMs as “mini-CEOs” finally comes to fruition.” With AI, if we still need to think in requirements, PMs will not be great, engineers will be better. If we don’t need requirements, business does not need product. Unclear where this is going, but I and many predict the next step will be product engineers, fusing product and engineering. My website is ugly because I made it (3 minute read) In the age of AI, it’s even more important to stick out, have a unique brand, perhaps an ugly website when everyone has the same slick LLM generated one. “Soon it will become something else entirely. Because it’s my website and I’m perpetually becoming somebody else.” Be yourself, be free, do not care what others say. ❤️ https://goodinternetmagazine.com/my-website-is-ugly-because-i-made-it/ The Who Cares Era (6 minute read) A glimpse into understanding the world: “It’s so emblematic of the moment we’re in, the Who Cares Era, where completely disposable things are shoddily produced for people to mostly ignore.” But then of course the article is wrong, talking about a large and expensive production: “He’s right. Nobody’s funding that kind of work right now, because nobody cares.” Nobody ever cared. It was just the elites, who cared about those things, were in control. They controlled the TV stations, and the NYT. But they no longer are. With the internet and social media they became irrelevant as gatekeepers and controllers. As the UK sitcom “Yes, Minister” (watch it!) said in the 1980s, why not subsidize football? We don’t but we fund opera houses, ballet and museums for the people who care about those things - https://www.youtube.com/watch?v=Zl0aEz34A4o - the masses finance the enjoyment of the elites because “culture!”. I agree of course with the ending, “Be yourself. Be imperfect. Be human. Care.” But don’t fool yourself, no one ever cared https://dansinker.com/posts/2025-05-23-who-cares/ AI: Accelerated Incompetence (8 minute read) LLM up, IQ down. The argument of the article. In previous news, calculator up, IQ down. On a stone tablet somewhere in Mesopotamia, Abacus up, IQ down. So nothing new there. “AI can degrade the quality of your codebase so fast.” Yes, you also can cut yourself with a knife pretty badly opening an Avocado - I’ve heard! If you don’t know your tools, don’t use them (But I love Avocado!) https://www.slater.dev/accelerated-incompetence/ Claude Code does our releases now (4 minute read) Something I’ve wondered about for some time now, are prompts the new open source frameworks? (article shares their prompt for the releases) https://www.aluxian.com/claude-code-does-our-releases-now/ The AI isn’t going to be on call at 2 AM when things go down. (3 minute read) Of course I do not agree with most of the article e.g. I think AIs are much better at understanding broad domains (maybe not narrow ones that only exist in your company), are much better at error handling, much better at creating secure code and so on, because as someone lately told me, “an AI is not bored by this”. Developers are. But I haven’t read this spark of an idea before anywhere else: “There’s a strange effect happening in teams using AI to generate code: nobody feels responsible for it.” Perhaps that breaks AI as it broke MDA and all the other developer replacement technologies before (it didn’t brake compilers, and from time to time I had to hear: “This must be a bug in the JVM!”) https://www.ivanturkovic.com/2025/05/22/llm-generated-code-liability/ Designing type inference for high quality type errors (28 minute read) Interesting article if you into this kind of thing (I am!). One pet interest of mine, compiler error messages and the article argues, type inference design influences the possibilites of describing type errors to the developer. Splendid! Not going to bore you (go read it if your interested as I am), here is a nugget I never thought of: Why is Go so fast in compilation? In the example, e.g. Java, the compiler needs to look
for the right https://blog.polybdenum.com/2025/02/14/designing-type-inference-for-high-quality-type-errors.html Management = Bullshit (2 minute read) “Enter the LLM. It can generate plans like nobody’s business. It can generate a plan for what to do in case of a fire, a meteor strike, or a zombie apocalypse. The plans are useless, naturally. They are just bullshit. But they satisfy management’s jonesing for plans, and best of all, they require no work on my part. It saved me hours of work yesterday.” 😀 Not endorsing that, in real it is helpful to think about these things. BUT in many large enterprises there is checklist security and disaster recovery theatre, no one cares if the plans work, they just need to be there somewhere, so .. https://funcall.blogspot.com/2025/05/management-bullshit.html When a team is too big (18 minute read) “Lately I’ve been thinking a lot about these questions: What is a team?” I know this, I know this, I know this! A team is a group of people who work together with a common goal. Everyone having different goals is not a team. “Many yeas ago I was part of a large team. 14 members to be exact. The problems started from the first standup: I didn’t follow 90% of what everyone else was talking about” I’ve been telling people for years, if a team meeting is boring 90% of the people all the time, you have the wrong team composition. Your team is not a team at all, it’s just a group of random people put into a “team” to organize and control them easier. One reason I prefer 2 devs + 1 pm as a team. But the article goes much deeper into the topic of teams and team size and composition. Most people I’ve met do not think enough about this. https://blog.alexewerlof.com/p/when-a-team-is-too-big We Did the Math on AI’s Energy Footprint. Here’s the Story You Haven’t Heard. (42 minute read) A good read, for a day with AI “You’d use about 2.9 kilowatt-hours of electricity—enough to ride over 100 miles on an e-bike (or around 10 miles in the average electric vehicle) or run the microwave for over three and a half hours.”. (on a sunny day our small balcony solar panels create roughly that amount of energy) But then measuring energy usage of AI today is like talking about efficieny of the first installed Diesel engine in 1898 (in my hometown of Kempten ;-) Yes relevant, but no not relevant. As shown in other articles shared in this newsletter, the amount of parameters to achieve one specific task goes down and I’m sure running AIs on essentially gaming graphics cards is great, but not the end of it. More on that in “The Poverty of Historicism” by Popper, one of my favorite ideas. https://www.technologyreview.com/2025/05/20/1116327/ai-energy-usage-climate-footprint-big-tech/ Join the CTO newsletter! | |
