If you only read one thingYouâve only added two lines - why did that take two days! (41 minute read) A classic. Why did it take so long? Because âthe issue was reported with a vague description of how to recreate it.â and âBecause the reported issue was related to functionality, Iâm not familiar withâ and âBecause I took the time to investigate the real cause of the issueâ and âBecause I investigated if there were other ways of getting to the same problemâ and âBecause I took the time to verify if there were other parts of the code that might be affected in similar ways.â and âBecause when I found the cause of the issue, I looked to find the simplest way of fixing itâ and âBecause I tested the change thoroughlyâ. Yes, dear CEO, I know you have no clue about software engineering, first thanks for asking, second thatâs why Iâm here. https://www.mrlacey.com/2020/07/youve-only-added-two-lines-why-did-that.html Video of the weekCapt. Grace Hopper on Future Possibilities: Data, Hardware, Software, and People (1982) (21 minute read) Great talk, 1982 (!), many nuggets, âWe have a very bad tendency to base our plans for computers on the equipment we have inhouse and the things weâre doing nowâ The biggest problem I see when I read tech visions of CTOs. Visions are for the future, the things âyou will be doing,â not the things youâre doing or going to do next quarter (thatâs OKRs for you, not a vision). If you, the CTO, are not a tech visionary, who in the company can be one? Work on the vision, if you need help, Iâm there for you. Another one: âOne of the major difficulties is the difficulty of changing peoples mindsâ Watch it. https://www.youtube.com/watch?v=si9iqF5uTFk
Stories Iâve enjoyed this weekđ Kobayashi Maru Management (19 minute read) âA Kobayashi erupts immediately. The swift response starts with someone raising their hand virtually or otherwise and what they say or type immediately differentiates this situation from your unexpected daily developments. You think, but do not say, âOh. Shit.ââ Even better, I like the article because how it suggests implementing change. Go read it! https://randsinrepose.com/archives/kobayashi-maru-management/ đ§ Founder Mode (8 minute read) Everyone talked about this one. Itâs about a talk to founders, by AirBnB founder Brian Chesky (note: AirBnB did a lot of awful things, but this is another topic about VC money). Now me, what can I add? First, Iâm a great admirer of Steve Jobs. As a founder (and CTO) you should run the company as he did (except perhaps be more polite). You decide what is a good product. Youâre the visionary, not some product guy youâve hired. Back to the article, founders get misleading advice because âwhat they were being told was how to run a company you hadnât foundedâ But then, not everyone is Steve Jobs, even if they think they are. Founder mode needs excellent people, âmanager modeâ just needs average people. The mistake founders made when loving the talk they listened to. They thought they are Steve Jobs and feel vindicated (also to be frank, AirBnB isnât the Steve-Jobs-era Apple either, neither is it Pixar or NeXT, itâs a private-hotel-website, sorry to digress again). But some nuggets, ââSkip-levelâ meetings will become the norm instead of a practice so unusual that thereâs a name for it.â Iâve heard about that idea the article talks about some years ago by one of the best CTOs there are, Christian von Hardenberg. If I recall correctly, in a talk at a CTO conference he said he doesnât manage in layers. He jumps up and down the hierarchy, sometimes confusing engineers why the group CTO shows up in an architecture meeting. He shows up where he thinks he can contribute crucial things. Since then, I tell my clients about this mode. Now Paul Graham some years later writes about it. https://paulgraham.com/foundermode.html No âHelloâ, No âQuick Callâ, and no Meetings Without an Agenda (13 minute read) Good tips how not to work asynchronously and remotely and what to do instead. No âHelloâ, then waiting for attention, just ask the question, this gets rid of the handshake. Also: CONTEXT! What did you do, what did you expect to happen, what happened instead. No âquick callâ (only when really necessary, like website down), âquick callâ just tries to pull asynchronous interaction into a synchronous mode. It tries to safe your time at the expense of your coworker or report. And yes, donât go to meetings without an agenda. Neither should your reports. https://switowski.com/blog/no-hello-no-quick-call-no-agendaless-meetings/ The Pragmatic Engineer Test: 12 Questions on Engineering Culture (13 minute read) There was the Joel Test https://www.joelonsoftware.com/2000/08/09/the-joel-test-12-steps-to-better-code/. Mucho important back in the days, then forgotten (sadly, as is Joel). I like the new engineering culture test. And the results of the survey! How do you score? https://blog.pragmaticengineer.com/pragmatic-engineer-test/ Startup CTO - Premature Scaling (3 minute read) Donât. No, donât. Seen this over and over again, âthat your much anticipated launch will be greeted by a vast, echoey silenceâ, CTOs anticipating a wave that doesnât come. And when itâs coming, always in surprising ways. Premature scaling not only costs time and money, it makes your app inflexible, complex and rigid. Then it will break if you need to pivot or adapt to the market. So, donât. https://blog.karliner.net/posts/startup-cto-premature-scaling/ đ˘ The U.S. Navyâs $100 Million Checkbox (29 minute read) A destroyer crashed into another ship. The reason? A checkbox in a graphical UI wasnât checked that coupled two ship drives. So the ship was unsteerable. $100M lost. Reminds me of that nuclear missle attack alarm UI, where the test button was next to the real button, and therefor a nuclear attack warning for Hawaii went out and people thought their lives would end because someone misclicked. UI! https://adrian3.com/blog/2019/2019-09-28-The-US-Navys-100-million-dollar-checkbox.php Ask for Advice, Not Permission (5 minute read) One of the biggest mistakes from engineering managers and developers moving into the CTO role. Youâre there to make things happen on your own, not ask permission. Once a CEO told me, âOnly ask my permission if your action could ruin the companyâ https://boz.com/articles/advice-not-permission Stonebraker Seeks to Invert the Computing Paradigm with DBOS (18 minute read) What if the database was the OS? Iâm still unconformable with that idea, but I have been raised that way as an engineer. Perhaps others will embrace the idea, and itâs the future. Some of my clients do REST and GraphQL directly from Postgres. Radical Simplicity indeed. https://www.datanami.com/2024/03/12/stonebraker-seeks-to-invert-the-computing-paradigm-with-dbos/ How to Hire a CTO (20 minute read) Not that you need to hire a CTO. Or you might, in the form of a successor. But in the end, the article is interesting for how to get that next CTO job. Another nugget in a newsletter of nuggets, âIncreasingly (~15% of the time) we see CEO/founders looking to delegate all of product development to one executive who needs to be skilled balancing customer relationships and what needs to be built (prioritization) with how the product is built (engineering). The Chief Technology & Product Officer (CTPO)â Many have tried, many have failed, I still believe itâs the way to go. So it seems does 15% of the market. https://www.khoslaventures.com/how-to-hire-a-cto/ How Superhuman Built an Engine to Find Product Market Fit (29 minute read) Many of my clients skip product market fit (PMF), then try grabbing money in opportunistic mode, run for more financing, waste money on marketing, fire people. Just to avoid working on PMF. PMF is hard, I get it. Working on PMF is hard work, slow progress, no grand features. Spending money on ad campaigns is more fun. Or going to startup events. Also people confuse PMF with traction (channel market fit). There are many frameworks for PMF, Superhuman used âWhat if we take our product away from you?â Work on PMF, donât skip it, donât work on the easy stuff. Donât pivot. https://review.firstround.com/how-superhuman-built-an-engine-to-find-product-market-fit/ Anthropic publishes the âsystem promptsâ that make Claude tick (21 minute read) Want to know what a âsystem promptâ is? How it affects your interaction with every AI? There you go. https://techcrunch.com/2024/08/26/anthropic-publishes-the-system-prompt-that-makes-claude-tick/ Challenges in Hiring and Growing Juniors in the age of LLM (13 minute read) Itâs important to have a mix of juniors and seniors. I see companies with too many seniors (one senior, ten juniors), and companies with too many seniors (where does the next wave come from when the seniors move on??) So juniors are important But ouch! âAre juniors still available? For my experience interviewing, no they are not. Even when the organization wants to hire them, they are so below the bar that the organization has set for itself, that those juniors cannot be hired.â I hire juniors: 1. They need to be able to code, I canât teach them that (and they might not even be able to code with the best teaching, ever!) 2. They need to have joy and discipline (aka motivation) 3. They need to be able to own things and feel responsible. Those will grow fast and become excellent. https://slowtechred.substack.com/p/software-engineering-in-the-age-of đ BooksWhat Iâm currently working on - Leave a price for me to decide!
Join the CTO newsletter! | |
