Amazing CTO | More happiness and success
🚀 129.4

by Stephan Schmidt

Happy 🌞 Sunday,

Welcome to my opinionated newsletter. This week’s insights

🚀 Is Software Really the UFOlogy of Engineering?
📈 AI’s Dial-Up Era: Boon or Bubble?
🔨 Why Rapid Prototyping is the CTO’s Secret Weapon
🌀 Scrum, Kubernetes, and the Myth of the Universal CTO
🧠 Simplifying Code: Functional Core or Ideological War?
🔎 Coding as a CTO: Bug Hunts and Brain Maps
🕵️‍♂️ How Senior Engineers Unexpectedly Lose Trust
🏷️ UUIDs: When Primary Keys Leak More Than Data

Good reading, have a nice Sunday ❤️ and a great week,

Stephan
CTO-Coach and CTO-veteran

Need support as an engineering manager? Thought about coaching? Let's talk—I helped many CTOs and engineering leaders with growth and making the right decisions under pressure, I can help you too.

🎁

If you only read one thing

Why prototyping matters (4 minute read)

(1) I agree with the article about the importance of prototyping. (2) It is great (3) I disagree with the methodology, in software a prototype is not “Can this work? Will it hold up? Does anyone care? Does it fulfill the desired purpose?” A prototype is to visualize the idea and discuss the idea and if the idea can ever work. The MVP is for “Does anyone care” and the PMF is “Does it fulfill the desired purpose”. ❤️ the “not the same pot” hint for the article’s prototype, MVP, MMP (Market ready - perhaps what I would call PMF) = don’t reuse the prototype, don’t scale the prototype: The elephant in the room, “In software, the temptation to “just tweak the prototype” is strong. After all, code is flexible—why not reuse it?” Indeed, why not? DON’T. “But that’s a trap.” I disagree with the article in details, but it’s great.

https://medium.com/@MarlonSchultz/why-prototyping-matters-fb7e0a5d5fae

🚀

Stories I’ve enjoyed this week

Is Software The UFOlogy of Engineering Disciplines? (13 minute read)

Long interesting article on UFOs - on the science, the art of science behind UFOlogy, or the non-science - on from there to the science of software. You can believe in aliens, when there is no, zero, zilch, nada evidence for aliens - but it is not science. Back to software, “Most empirical studies into our discipline are small, attempting to extract meaningful trends from statistically insignificant sample sizes. This leaves them wide open to statistical noise.” which is the sad state of our industry. “If software engineering ‘s to be truly scientific, our hypotheses need to be refutable.” Why is this not happening? Because there is no need. Because software development works beautifully without science. It makes billions upon billions of dollars. In other areas, science has helped the industry make more money - or achieve things that were not possible before. In software there is so much money that science is not needed (except in some fundamental cases, like data structures). But, sad state. My belief, not science, many companies could save lots of money when using Go instead of Python (you might disagree - love you for that). Software is such a great multiplier, that it’s not important if you use Python or Go. It’s important you use software. CEOs on one hand want to know if software engineering is efficiently working on the right things, while at the same time, their whole company would not work without software. Therefore, we don’t care if software engineering is scientific or not. And as a result we’re driven by ideologies, fashion and hypes.

https://codemanship.wordpress.com/2025/11/07/is-software-the-ufology-of-engineering-disciplines/

AI’s Dial-Up Era (16 minute read)

Does AI create more jobs or less jobs? The article cites Jevons Paradox, something you hear often today. It says when something becomes more efficient, we will get more of it, not less. “When we make a technology more efficient, demand goes well beyond the original level.” The challenge is that a lot of AI is geared towards production of things we need to consume (media and apps) - and people are at peak consumption, at least in the rich countries. The biggest competitor to new AAA videogames are not AAA games launching at the same time, but Fortnite and TikTok. People are at max consumption. Again the article has some nice graphs and some data to its claims. “Once demand saturates, employment doesn’t further increase but holds steady at peak demand. But as automation continues and workers keep getting more productive, employment starts to decline.” but “A restaurant owner might use AI to create custom supply chain software that with human developers, say at $100,000, would never have been built.” Or as I posted several times over the last weeks, 1-person-apps, applications for an audience of 1. Do they create jobs? Is this where the productivity and lower costs of software development go? Not more developers but more creators?

https://www.wreflection.com/p/ai-dial-up-era

The kind of company I want to be a part of… (2 minute read)

A small difference, ‘1 apple’ and ‘2 apples’ vs. ‘1 apple(s)’. The author wants to live in the first kind of company. I call this being proud of what you create. I call this engineering craftsmanship. That’s also my company.

https://www.dvsj.in/my-company

Unexpected Things that are People (7 minute read)

“Ships are accorded limited legal person rights, primarily so that they can be impounded and their property seized if they do something wrong.” Didn’t know that one. Boom. Worldview changed. Many more interesting examples in the post. Will AI be accorded limited legal person rights? Like ships?

https://bengoldhaber.substack.com/p/unexpected-things-that-are-people

dead framework theory (14 minute read)

“React isn’t competing with other frameworks anymore. React has become the platform. And if you’re building a new framework, library or browser feature today, you need to understand that you’re not just competing with React—you’re competing against a self-reinforcing feedback loop between LLM training data, system prompts, and developer output that makes displacing React functionally impossible.” I do think this is transitionary and frameworks and libraries will go mostly away. Article has lots of nice graphs.

https://aifoc.us/dead-framework-theory/

Claude Code Can Debug Low-level Cryptography (8 minute read)

How good are AIs? How good is Claude Code? It can find bugs in crypto code, “To my surprise, it pinged me a few minutes later with a complete fix.” I’ve seen this myself: Claude Code is astonishingly good at finding bugs. Because while people are talking about AI context windows, and how small they are, my context window is 7 things and Claude beats me hands down. It’s faster and better at finding bugs. Bugs that are based on explicit ideas in the code. It’s bad at everything magic, implicit things like ORMs - with everything ‘Convention Over Configuration’. For my website which is created with Hugo, where implicit rules determine which templates are used in which order and context, Claude Code struggles. For AI, explicit is better. And when the bug is somewhere in a dozen files of code, Claude Code will be better at finding it than you.

https://words.filippo.io/claude-debugging/

Active Listening: Swiss Army Knife of Communication (with Examples) (27 minute read)

“Active listening is a way to understand someone’s point of view. While you listen, you ignore your own thoughts and ideas.” Super power - but depending on your personality, very hard work. Worth to practice though, every day.

https://togetherlondon.com/insights/active-listening-swiss-army-knife

What the hell is a CTO? (7 minute read)

“Like with planning, these are tricky discussions to navigate because a) nobody really agrees on what the hell a CTO is and b) even if we did, it’s so company — and company stage — dependent that the agreement would be an illusion. “ He’s right. I’ve been thinking about a CTO podcast called “Why I’m different” because I think we try to make every #CTO too much the same, same 2-Pizza-Teams, same Tribes, same Scrum, same Kubernetes - when in reality every CTO job is very different. It’s perhaps this that I’ve learned from my CTO coaching over the last 8 years. Surprisingly well reflected and well written article, with “Stay too close to the tech, too close to all the critical decisions, and you deprive your company and teams from the chance to grow as leaders and technologists. [..] drift too far away and your team — and CEO — loses a critical voice and thought partner. “ If longer, would have made the “Read one thing” category.

https://cory.news/posts/2025-10-31-cto/

Simplify Your Code: Functional Core, Imperative Shell (5 minute read)

One thing I dislike a lot in our industry, we are driven by ideology. One of those ideologies is “functional programming” (FP) - not as a term to describe a way to structure code, but as a holy grail to achieve. What is functional programming? Whatever we want it to be it seems. But functional programming is simple: Programming where functions take other functions as parameters and (can) return functions as results. func security(f func(int) string) func(int) string That’s FP. Everything else is not exclusive to FP. Like side effect free functions, data first, or immutable data. Good developers in any language - even Java ones - will use side effect free functions as much as possible, and will use immutable data as much as possible. In the article function getExpiredUsers(users: User[], cutoff: Date): User[] is given as functional code. It does not take a function as an argument, it does not return a function. The one thing: it is side effect free. Which does not make this functional code by itself. It rains, therefore the ground is wet. The ground is wet does not mean it rained. Elementary My Dear Watson!

https://testing.googleblog.com/2025/10/simplify-your-code-functional-core.html

Why I code as a CTO (8 minute read)

Some good ideas on why to code as CTO. I do agree with some of them, not with the “A customer wants something urgently” though - you need to change the process not cover the gaps. Found this one interesting: “People are often shocked by this, but I fix a lot of bugs! And bugfixing is one of my favorite ways to maintain a mental map of our codebase.” Yes, keep coding if you’ve done everything else expected of you (things only you can do that is!)

https://www.assembled.com/blog/why-i-code-as-a-cto

How Senior Engineers Lose Trust (2 minute read)

People get promoted, but “Same behavior. Different expectations.” I’ve seen this with developers becoming managers all the time. They still act like ICs and don’t own parts of the company problems. You’re no longer a developer, act like a manager now. Key to being a manager is owning problems and thinking about the box you’re in. As a manager you’re expected to change the box, the rules you can change.

https://tahahussain.substack.com/p/how-senior-engineers-lose-trust

Primary keys using UUID v7 are (potentially) an HR violation | Casual Pontifications (3 minute read)

UUID7 are the new UUID on the block and they have a time based component. The time part leaks when the key was created - and depending on what it represents, e.g. a job application id or a patient id, it leaks confidential information. It’s not new, it’s already there with other IDs, but this one will bite many CTOs in the future. Be careful what information you leak.

https://mikenotthepope.com/primary-keys-using-uuid-v7-are-potentially-an-hr-violation/

What is Stephan doing?

More 1-person-apps: I’ve (Claude) written an app that creates diagrams for my Theory of Control. From text representations like

1,1: OKRs
>+1,+1: Business Initiatives
>+1,+1: Backlog Planning, rb

it will draw a diagram. Without AI this would have been too much effort with the layout algorithms and future maintenance. Now it’s easy and saves me a lot of time (Small tip, let #ClaudeCode use inkscape to convert SVGs to PNGs to be able to see them - another problem of AIs, they have trouble fixing visual bugs if they don’t see them

add the inkscape hint in CLAUDE.md).

I was also giving a talk on “AI is not software” at code.talks in Hamburg. Excellent event, excellent team. Got great feedback afterwards. The talk is about how AI will replace software as a tool. Also did some podcast recordings as a guest for different podcasts, which will be published next year. Wave of new CTO coaching clients, busy times overall.

Events for Engineering Managers