Blog

I spent years of my career getting paid well to patch applications at companies that had refused to invest in their engineering teams. The pattern was always the same, and I am watching it play out again with AI. All of this stemmed from the company’s unwillingness to invest in the skills of their team and in some cases force engineers to learn new technologies. It is often the case that the company itself, especially at the senior executive level does not care what tech stack is used as long as the application runs well and the company can sell its products and services. There might be some directive like using a certain vendor for ecommerce but that rarely prevents day-to-day operations shifting with industry trends. Instead the cause of these issues would be engineers and managers of those teams finding some reason to dislike SPAs and newer web technologies, often irrationally, then once that happens for a couple of years it becomes very hard to jump back onto the train. Eventually it becomes harder to hire, your most excited forward-thinking team members leave and you are stuck with a decaying application hiring contractors to patch holes. ...

You will not have forever to learn the fundamentals of AI engineering and by choosing to ignore this class of technologies you will fall behind in your career like those who did not learn React, Vue and never developed single page application (SPA) knowledge. Every week I come onto Discord, Slack or X and see new tools for AI engineering and AI-assisted development that solves some gap previous tools struggled with. Some of these tools are developed in-house based on a specific organization’s need and some are publicly available but what is important is that you recognize these tools are being released frequently and each are helping direct the way our industry moves forward. Choosing to ignore the changes to our industry will not slow the change or keep your skillset relevant for longer. ...

Tickets are the most common unit of work I use as a software engineer, they often come from a product manager, program manager or even other engineers. In my time being a contractor, engineering manager and full time team member I have seen several styles for writing tickets not all of them good. This post is my attempt to write what I believe to be good practice when writing tickets and the minimum fields required in the hopes that your journey in software engineering is efficient, tickets are not wasted and your work remains high quality. ...

For several years I have had to deal with repeated emails from my old cPanel host telling me that so-and-so plugin is out of date or that some critical vulnerability has been found in WordPress. Frankly, I have been sick of using WordPress for my personal website for a long time. I originally set it up using cPanel on VentraIP because it was cheap ($12/year) and easy to manage especially with auto-updating plugins and themes but I have been considering migrating to something simpler and easier to use for at least 5 years. ...

Building for the K-shaped economy is something I think all founders need to start considering. The K-shaped economy refers to a pattern we are seeing in the US and other western countries where we have some people doing very well. These are people working in tech, finance, law, and even trades earning more money than they ever have particularly for business owners. The other branch of the K we have people performing worse over time with lower wage growth and lower material conditions. The advent of AI has not helped this in some of those same high performing industries with thousands being laid off from tech companies, automation like self-checkouts removing roles for young people and less skilled staff. Computer science students, once all but guaranteed high paying jobs are now struggling to find work. ...

An interesting topic arose over the holiday period which is people using AI tools for the first time or at least first time in a while and realizing how far they have come since last year. Claude code in particular had new users seemingly everywhere. The strange thing about this for the AI maxi crowd is that we have been using these tools all year. Some of the reviews have been panicked as people realize just how far some of this tooling has come in the last 12 months. ...

I spent 2025 telling myself I was too busy to write. I wasn’t too busy, I just stopped trying. It has been a while since I last updated this blog with zero posts in 2025. This is not something I planned or expected but as with most people I simply “got busy” and had other things on my plate. I am now reviving this site and plan to write at least one post per week covering anything across my usual topics of culture, tech and management. ...

A couple of weeks back, I found myself in a discussion about ‘dark matter developers’ - a term coined by Scott Hanselman to describe the vast majority of developers who, like dark matter in physics, we know exist but rarely see or interact with online. The discussion of these developers started because the Cursor AI team were shocked to find that 80% of their users are on Windows with Laurie Voss tweeting about the episode: “Every so often Silicon Valley devs re-discover that the vast majority of programmers use Windows and are re-surprised.” I responded noting how minds are blown when people realise most developers don’t work at tech companies but in traditional company IT departments. ...

After ten years the version 5 pull request for Express has been merged. The original pull request was created back in July 2014 and from speaking to a lot of software engineers had been considered dead. This release is intentionally boring with the focus being security enhancements and improvements to the core foundations to enable faster development in the future. This has parallels to Node prior to the more frequent release cadence after being pushed by iojs. Express v5 will only support newer versions of Node, with the oldest supported version being 18 (maintenance LTS, released 2022). Security concerns and refinement of patterns are core to the release as is simplifying maintenance. ...

I had a whole post ready for today to discuss “dark matter developers” and the silent majority that keeps our software applications running. After reading Anthropic’s press release for new Claude models today I felt that was more important and the dark matter developers post will be up next week ✌️ ​ ​Claude 3.5 Release​ ​ Anthropic has released an updated Claude 3.5 Sonnet model, showing major improvements in software development capabilities. The new model scored 49% on SWE Bench tests (up from 33.4%), Claude was already the best model for software development and it has now extended the lead scoring higher than all other publicly available models in solving real-world GitHub issues in Python repositories. ​ Claude Sonnet’s understanding of business domains has also improved with performance on the TAU-Bench metrics increasing from 62.5% to 69.2%. This means that Sonnet is more capable of reasoning and understanding compared to previous models and most importantly when compared to other public models. ​ Claude 3.5 Haiku has some updates as well. It is now more efficient and has the same cost as the previous generation of Claude Haiku 3.0 with much better results. Haiku can now outperform the previous generation of Opus and remains the best option when building user-facing AI applications. ​ Interestingly, Claude Opus has been removed from the model list where it was previously marked as “Coming Soon”. Online, people are speculating this is because Anthropic is able to get better results with a smaller model, if so that is very exciting and continues to show the growth in research and increasing power of AI systems. ​ The most exciting part of this release is the Computer Use feature where Claude can now take on the role of a user in a real world application accessing browser windows and even the computer itself. In their demo video Anthropic researchers show Claude filling out a form based on spreadsheet data, other online information and information from the user’s computer. The possibilities of this to increase productivity is massive with many data entry tasks being fully automated. ​ One thing I do wish Anthropic would resolve is their versioning, it would be much better if this was called Claude Sonnet 3.6 instead of reusing the 3.5 version number. This duplication of versioning means it is more difficult to understand what people are referring to particularly in blogs or on X. ​ If you are curious about how Generative AI and LLMs can help your business particularly if you are an engineering manager or software engineer please reach out. I would also love to hear from you if you have interesting use cases or examples from your real world workflows. ​ If you are interested in following my journey you can subscribe to my blog and YouTube channel. My other links are available on My Linktree.