Month: August 2025

• August is the last month of my employer’s financial year, which means preparing for annual appraisals just as everyone is using up their leave – and this is on top of client work. While I’m enjoying being busy, it’s also been tiring.
• We finished off our internal Summer of Q, where about 40 of us explored agentic AIs for coding. It’s one of several things I need to write up before I forget about them. I love working with agentic AI, but doing it effectively will take some more work.
• I continue to think about Sean Goedecke’s essays on value, alongside his one on Pure and Impure Software Engineering. I’m mentioning Goedecke in most of my weeknotes recently. I have been rereading some of his pieces and want to put them into practise when I start a new role in September.
• My new role involves React, so it’s time to get to grips with that properly.

Links

• From Noah Brier’s Things I think about AI: “If models didn’t progress from what we have today, we’d still have 10 years of runway to integrate their value into corporations and society.”
• I’m not convinced that about the power of agentic coding with tools like Kiro and AmazonQ. But I do think there is a use for small, focussed tools. Bill Mill gives a good example of this with a code review tool.
• Two good checklists on agentic coding techniques:
  • AI Gives Time, Not Confidence: Developer Productivity Toolkit by Jonathan Vila
  • Coding with LLMs in the summer of 2025 (an update)

• It’s been a few weeks since my last weeknotes. I had a restful holiday in Wales and a lot of work things to catch up on.
• I’ve been thinking a lot about Sean Godecke’s essays on the role of a programmer in the current economy, and the need to provide demonstrable value – particularly for seniors and managers. This is shaping my approach to my work.
• I wrote a post about how The Joel Test turned 25 this month, looking at its historical importance and pondering what a modern Joel test might include.
• In another flight of nostalgia I looked up Brad Fitzpatrick and Lisa Philip’s notes on scaling Livejournal, which were a useful reference for scaling Flirtomatic many years ago.
• I continued my experiments with Amazon Q and learned a lot. I should have those written up by the end of the month.

Links

• The 2025 Stack Overflow Developer Survey has been published, although it may be a few weeks before I get around to reading this properly.
• Another great piece from Charity Majors was In Praise of ‘Normal‘ Engineers, where she talks about the importance of teams rather than individual developers, and how systems must be easy to use: “The best engineering orgs are the ones where normal engineers can do great work”
• Sean Goedecke’s What’s Going to Happen to Junior Engineers was a good exploration of the potential long-term meaning of the fall in junior tech positions.
• The conclusions of Anthropic’s Project Vend feel quite speculative, but it raises some interesting questions: “we think this experiment suggests that AI middle-managers are plausibly on the horizon… the AI won’t have to be perfect to be adopted; it will just have to be competitive with human performance at a lower cost in some cases… we don’t know if AI middle managers would actually replace many existing jobs or instead spawn a new category of businesses”.
• Welcoming the Next Generation of Programmers argues that vibe coders are programmers and it’s important to onboard and welcome them to the existing developer communities.
• Christina Wodtke posted on linkedin about how a lot of Gen-X programmers have the same passion for GenAI as for the early web – in a way that didn’t happen with blockchain or VR (via Simon Willison)

Books

Tidy First by Kent Beck

I read Kent Beck’s Tidy First? while away – some thought-provoking ideas, but they rely on being able to produce and merge changes easily. Again, I need to find time to write up my notes.