https://simonking.dev Notes on software, leadership and engineering judgement. en-GB https://simonking.dev/writing/architecture-trade-offs-not-just-patterns Well-patterned systems and well-architected ones are not the same thing. Every architectural decision gives something and takes something away - the teams that build systems that hold up can explain both sides. Thu, 11 Jun 2026 00:00:00 +0000 https://simonking.dev/writing/architecture-trade-offs-not-just-patterns https://simonking.dev/writing/prevention-is-invisible When you stop a problem before it exists there is no incident, no war story, no late night. Just a system that kept working. Engineering cultures that only reward visible problems quietly build the ones they can't see. Mon, 08 Jun 2026 00:00:00 +0000 https://simonking.dev/writing/prevention-is-invisible https://simonking.dev/writing/team-balance-over-alignment Teams where everyone thinks the same way feel efficient at first. The gaps show later. The strongest teams have intentional friction - people challenging assumptions and forcing ideas to hold up before they become decisions. Sat, 06 Jun 2026 00:00:00 +0000 https://simonking.dev/writing/team-balance-over-alignment https://simonking.dev/writing/optimise-for-the-problem Knowing the correct approach matters, but the job is to solve the right problem for the right person in a way that actually works. The gap between a good engineer and an effective one is knowing when to follow the theory and when to set it aside. Fri, 05 Jun 2026 00:00:00 +0000 https://simonking.dev/writing/optimise-for-the-problem https://simonking.dev/writing/technical-debt-is-a-business-problem Raising technical debt as "time to clean things up" rarely lands. The teams that handle it well don't talk about debt - they talk about what it's costing the business right now. Thu, 28 May 2026 00:00:00 +0000 https://simonking.dev/writing/technical-debt-is-a-business-problem https://simonking.dev/writing/ai-changes-which-problems-are-worth-solving Every engineer has a mental list of tasks that aren't worth automating yet. AI doesn't just make you faster - it changes the threshold for what moves from "not yet" to "done". Wed, 27 May 2026 00:00:00 +0000 https://simonking.dev/writing/ai-changes-which-problems-are-worth-solving https://simonking.dev/writing/building-the-wrong-thing-well The worst mistake I made early in my career wasn't bad code. It was building something technically solid that nobody needed. The best technical work isn't impressive because of how it's built. It's impressive because of how precisely it fits the problem. Thu, 21 May 2026 00:00:00 +0000 https://simonking.dev/writing/building-the-wrong-thing-well https://simonking.dev/writing/pull-requests-as-signal Most teams say they care about quality. Pull requests are where you find out if that's true. AI makes PR discipline more important, not less. Tue, 19 May 2026 00:00:00 +0000 https://simonking.dev/writing/pull-requests-as-signal https://simonking.dev/writing/saying-yes-is-debt Saying yes keeps people happy and looks like responsiveness. But strong engineering teams don't just ask whether they can build something - they ask whether they should. Fri, 15 May 2026 00:00:00 +0000 https://simonking.dev/writing/saying-yes-is-debt https://simonking.dev/writing/decisions-cheap-to-revisit Engineering decisions don't reveal themselves in the moment. They show up when the system is under pressure. The goal isn't to be right - it's to stay recoverable. Fri, 15 May 2026 00:00:00 +0000 https://simonking.dev/writing/decisions-cheap-to-revisit https://simonking.dev/writing/the-most-interesting-thing-happening The enterprise AI fight stopped being about models a long time ago. OpenAI has mindshare but no enterprise packaging. Anthropic has brilliant packaging but a quality process that is starting to crack under the pace of expansion. Sat, 09 May 2026 00:00:00 +0000 https://simonking.dev/writing/the-most-interesting-thing-happening https://simonking.dev/writing/the-hardest-part-of-leading Effective engineering leadership isn't about stepping in and making calls - it's about knowing when to hold back, so the team builds the ownership and judgement that makes them stronger. Fri, 08 May 2026 00:00:00 +0000 https://simonking.dev/writing/the-hardest-part-of-leading https://simonking.dev/writing/architecture-decisions-cost The most expensive decisions in software aren't about tools or people. They're about architecture - and you don't pay for them once. You pay for them every time the system changes. Tue, 05 May 2026 00:00:00 +0000 https://simonking.dev/writing/architecture-decisions-cost https://simonking.dev/writing/ai-tools-missed-the-brief I ran the same brief through ChatGPT and Claude. The real test was what I left unsaid. AI tools are powerful at execution - the judgement about what to build still sits with the engineer. Sat, 02 May 2026 00:00:00 +0000 https://simonking.dev/writing/ai-tools-missed-the-brief https://simonking.dev/writing/ai-speed-without-discipline AI makes the trade-off between speed and quality more visible. When time stops being the constraint, the decisions about what you ship and how carefully say more about the team than the tools. Thu, 30 Apr 2026 00:00:00 +0000 https://simonking.dev/writing/ai-speed-without-discipline https://simonking.dev/writing/production-cost-thinking Engineers who develop the sharpest instincts aren't the ones with the most resources. They're the ones who never had the option to ignore the bill. Tue, 28 Apr 2026 00:00:00 +0000 https://simonking.dev/writing/production-cost-thinking https://simonking.dev/writing/claude-design-not-a-production-tool Claude Design commits to a direction before you have had the chance to explore - that is a signal about what the tool is actually for, not a gap in capability. Sat, 25 Apr 2026 00:00:00 +0000 https://simonking.dev/writing/claude-design-not-a-production-tool https://simonking.dev/writing/ai-code-that-looks-correct AI presents working code and broken code in exactly the same way. That confidence problem is a leadership problem, not a tools problem. Thu, 23 Apr 2026 00:00:00 +0000 https://simonking.dev/writing/ai-code-that-looks-correct https://simonking.dev/writing/the-knowledge-moat When AI makes building cheaper, the advantage moves to whoever understands the problem deeply enough to build something that actually fits - that is a knowledge moat, and it is much harder to copy. Tue, 21 Apr 2026 00:00:00 +0000 https://simonking.dev/writing/the-knowledge-moat https://simonking.dev/writing/environments-you-couldnt-hide-in Engineers who develop judgement fastest are the ones who chose environments where consequences were visible, feedback was short, and there were no layers to absorb the impact. Sat, 18 Apr 2026 00:00:00 +0000 https://simonking.dev/writing/environments-you-couldnt-hide-in https://simonking.dev/writing/developers-not-expensive-decisions-are The salary debate misses the point. Good engineers reduce the cost of decisions - and that has always been where the real expense lives. Fri, 17 Apr 2026 00:00:00 +0000 https://simonking.dev/writing/developers-not-expensive-decisions-are https://simonking.dev/writing/developing-developers Once you understand that the real job is creating conditions for the team to do their best work, everything about how you spend your time changes. Sat, 11 Apr 2026 00:00:00 +0000 https://simonking.dev/writing/developing-developers https://simonking.dev/writing/nobody-tells-you-when-you-become-tech-lead The thing that got you the role is the thing you have to learn to step back from. The real job isn't writing code anymore - it's making sure good decisions compound over time. Fri, 10 Apr 2026 00:00:00 +0000 https://simonking.dev/writing/nobody-tells-you-when-you-become-tech-lead https://simonking.dev/writing/best-interview-question The best interview signal isn't on a CV - it's how someone responds when asked about a decision that turned out to be wrong. Thu, 09 Apr 2026 00:00:00 +0000 https://simonking.dev/writing/best-interview-question https://simonking.dev/writing/if-everyone-agrees-with-you The most valuable person on a team is often the one who tells you when you're wrong. Good engineers build cultures where disagreement is safe - and that matters more now than ever. Sat, 04 Apr 2026 00:00:00 +0000 https://simonking.dev/writing/if-everyone-agrees-with-you https://simonking.dev/writing/consistency-over-brilliance The same trade-off you'd make in hiring - consistent reliability over occasional brilliance - applies directly to how you choose and build with LLMs. Fri, 03 Apr 2026 00:00:00 +0000 https://simonking.dev/writing/consistency-over-brilliance https://simonking.dev/writing/smaller-focused-ai-model More capability isn't always better. In production, a smaller, more focused model you can understand and trust often outperforms a general-purpose one that's harder to control. Thu, 02 Apr 2026 00:00:00 +0000 https://simonking.dev/writing/smaller-focused-ai-model https://simonking.dev/writing/hiring-fewer-junior-developers Optimising for short-term delivery by not hiring juniors is a trade-off that doesn't show up on a sprint board - but it costs you the pipeline that builds your future team. Sat, 28 Mar 2026 00:00:00 +0000 https://simonking.dev/writing/hiring-fewer-junior-developers https://simonking.dev/writing/ai-thinking-not-just-coding For neurodivergent engineers in particular, AI's real value is in helping structure thinking and make communication easier. That's a more meaningful advantage than raw output speed. Wed, 25 Mar 2026 00:00:00 +0000 https://simonking.dev/writing/ai-thinking-not-just-coding https://simonking.dev/writing/quality-in-the-right-places Good engineering isn't about ignoring quality - it's about applying it where it actually improves the product. Time spent perfecting the wrong thing is still wasted. Tue, 24 Mar 2026 00:00:00 +0000 https://simonking.dev/writing/quality-in-the-right-places https://simonking.dev/writing/ai-tools-vs-ai-judgement Access to AI tools is becoming a commodity. The real advantage is the judgement behind how they're used - knowing when to use them, what to trust, and when not to build at all. Mon, 23 Mar 2026 00:00:00 +0000 https://simonking.dev/writing/ai-tools-vs-ai-judgement https://simonking.dev/writing/prototype-vs-production AI is making prototyping faster and that's genuinely useful. The cost of building something that can run in the real world is a separate question. Prototypes show the idea. They are not the thing. Sun, 22 Mar 2026 00:00:00 +0000 https://simonking.dev/writing/prototype-vs-production https://simonking.dev/writing/what-senior-engineer-means Early on, seniority looks like speed and output. Over time, you realise it's about judgement - knowing what's worth building in the first place. Wed, 18 Mar 2026 00:00:00 +0000 https://simonking.dev/writing/what-senior-engineer-means https://simonking.dev/writing/betting-google-wins-ai Most people frame AI as OpenAI vs Anthropic. But winning usually comes down to distribution, unit economics, and the ability to ship at scale - which points to Google. Sat, 14 Mar 2026 00:00:00 +0000 https://simonking.dev/writing/betting-google-wins-ai https://simonking.dev/writing/what-makes-a-good-dev-in-2026 The industry is splitting into camps around AI. The developers who thrive are the ones who can do both - ship and think - with range across fundamentals, judgement, and stakeholder fluency. Thu, 12 Mar 2026 00:00:00 +0000 https://simonking.dev/writing/what-makes-a-good-dev-in-2026