This weekās insights
Good reading, have a nice Sunday ā¤ļø and a great week, Stephan
If you only read one thing50 things weāve learned about building successful products (30 minute read) Perhaps the most important article this year up to date. So many true things, how to build successful products (or successfully products) - like āSmall teamsā, āhigh bar of hiringā, ātrustā, ātransparencyā, āwaiting to ship is badā, āAdopting new technologies into your product should only be done for hair-on-fire problemā and many many more. This one is to print out and read every day. I fell in ā¤ļø love with the article by āArtificial deadlines will not make your team faster.ā Everything in this article is gold. Again, I love people sharing experiences from the things they have done, vs. Agile Coaches who never were managers, are no coders, have no clue, got some training and then tell you what to do - for a lot of money. Listen to people who have been there, walked the walk, wore the shoes, not to people who saw it on TV. And read that great article. https://newsletter.posthog.com/p/50-things-weve-learned-about-building
Stories Iāve enjoyed this weekWith AI You Need to Think Much Bigger! (10 minute read) āI have noticed that I am no longer scared that a project will be too big or too complex for me, or that a project will use a technology or programming language I donāt knowā At the end of the authors career, AI āfeels like wining the lottery two weeks before you dieā. AI is not a quantitive change, itās a qualitive change. Repeat that every day you wake up. Act accordingly. Defining a Great Engineering Culture (40 minute read) I call engineering culture āautomatic managementā. People making the right decisions without you making them, or you handholding people, or you micro managing. For this to work donāt make the biggest culture mistake: Bottom up culture definition. Make it top down. The article has some good points like āStart with the company cultureā, indeed, donāt define something that goes against company culture or wants to build a separate company. Some ideas on what values you should have, donāt agree on all of them, also feels too fluffy from time to time. Engineering culture needs to be practical, āShould I do A or B?ā - thinking about the culture - āAh B of courseā. If your culture canāt do this, drop it. Overall the article is a long, but good one. https://mgrebler.substack.com/p/defining-a-great-engineering-culture No Co-Founder Needed: Being a Solo Founder (20 minute read) Interesting article. From being a founder in serveral companies, I think the downside of being a solo founder is your mood swings (depending on your personality). Startups are really tough, two people can carry each other, when youāre alone you need lots of discipline and motivation. Which brings me to the point of AI (ha, you thought I can smuggle AI into this one): With AI, are there teams in the future? Isnāt everyone working on their own? Are we not all solo founders in the AI future inside a company? https://eidel.io/no-co-founder-needed-being-a-solo-founder/ Apple does AI as Microsoft did mobile (3 minute read) Steve Ballmer famously laughed at the iPhone (āit has no keyboard!ā). Now John Ternus, SVP at Apple, laughs at AI. DHH writes: āBut Cook evidently does not have the product savvy to be able to tell bullshit from benefit, so he keeps giving Giannandrea [VP of ML/AI] more rope.ā Ah the famous discussion about the CEO of a product company. Just like Steve Jobs, you need to know what works and what doesnāt. Or as the CEO in Margin Call says āDo you know why I earn the big bucks? Iām here for one reason alone. Iām here to guess what the music might do one year from nowā or in DHHs words ātell bullshit from benefitā. And Cook, as most second-in-command, is good at optimizing, but has no vision and canāt distinguish good from bad products. And there is a linked interview where John Ternus laugs at AI. Steve Jobs would have never let anyone else speak about the vision of his company. This is why Apple is doomed, āThis is what having a company run by a logistics guy looks like.ā https://world.hey.com/dhh/apple-does-ai-as-microsoft-did-mobile-df2c98ca Y-Combinator: The First 20 Years (10 minute read) A book costing $250. And still Iām tempted to buy it. What again makes me wonder, who am I? And: Your product can be much more expensive than it currently is, if it has pull, people will pay (Also the rumored $20.000/month OpenAI licenses). https://ypaginator.com/store/p/thxyc Claude Code blew my mind several times this week. The only downside currently, it keeps adding features it thinks I need, but which I donāt need! And itās expensive. https://x.com/Steve_Yegge/status/1898674257808515242 Why Layoffs Donāt Work (10 minute read) There are good reasons for layoffs and bad reasons. Good reason: Youāre going bust otherwise. Cut early and deep. Bad reason: Increase profit. This leaves everyone in the company demotivated. Whenever you lay off people, take care of those you lay off, but double so for the people who stay. Article has some study data on performance of companies who laid of people for profit (not good). PS: Instead of laying off people, fix hiring and people management. https://thehustle.co/originals/why-layoffs-dont-work How to Write Useful Commit Messages (23 minute read) 1. Give this to all engineers, I have seen so many bad commit messages. Why have them at all then? Because your tool prevents them from being empty? Ask WHY you have commit messages in the first place, what should they accomplish 2. Isnāt an AI today much better at creating them? Or canāt this be done, because the AI canāt know your intent? Or perhaps it can from the changes? Me personally? As CTO I would use an AI for this, not push developers to write better messages. https://refactoringenglish.com/chapters/commit-messages/ You might not need Redis (7 minute read) I have put Redis everywhere, because it just works and does not break. BUT there is truth in the article, do you really need Redis? Or is it just a bandaid? Three examples in the article on how Redis was added but not needed. Perhaps you donāt need it either? What else donāt you need and have just introduced as a reflex or bandaid? https://www.viblo.se/posts/no-need-redis/ How did places like Bell Labs know how to ask the right questions? (43 minute read) The question of the internet age was, how do build the right things, not how to build as many things as possible. Itās going to be even more important in the AI age. Something CEOs still do not understand and push CTOs to deliver more faster, instead of focusing on impact and have a clear understanding about what to deliver. And CTOs are not involved enough in research and innovation, instead focusing on execution - which culminated in the last years on this strange (and misguided!) metrics fetisch in engineering. Instead, read the article and learn āhow to ask the right questionsā https://www.freaktakes.com/p/how-did-places-like-bell-labs-know Join the CTO newsletter! | |
