May 2026

There’s so much to do and so little time!

I’d like to:

• Get LaptimeInsights to MVP - getting it to it’s minimum useful state would be a good win. It’s not in bad shape at the moment, but it needs pulling together and testing to make sure its useable.
• Finish my OpenSpec experiment - I’m building an application using OpenSpec to see how it works. I’m hoping that once the specs are built I can try feeding these into other code harnesses like OpenCode with a local LLM etc and see what the difference is.
• I’ve been coding a small “example” library so I could refresh my memory on Java and SpringBoot. I’ve been using Kotlin and KTOR etc recently so I’m a little rusty. It would be good to finish this off and write about the experience - AI has produced code, but not quite how I want it. It’d be interesting to refactor it and make it what I want it to be, and then compare the two versions…
• Do some AI training!!! There’s a lot to know and catch up on, and there’s more every day.
• And I would like to do some YouTube videos for the above projects, but I need to skill up there!

Things I’m seeing/reading/watching/listening to in May 2026…

Videos

• “Software Fundamentals Matter More Than Ever” — Matt Pocock
  • Suggests that spec driven design isn’t the answer
• Can You Outsmart the Model Makers?
  • From the “AI that works” podcast: What is harness engineering, and how can you build harnesses that genuinely push AI performance? We cut through the noise, discussing whether a ‘dumb’ model with a great harness can outperform a ‘good’ model with a poor one.
• The New Code — Sean Grove, OpenAI
  • A presentation from the AI Engineer conference discussing how code is a lossy projection of intent, moving from ephemeral “vibe coding” prompts to durable, version-controlled markdown specifications that align both humans and AI weights.

Podcasts

• Stepchange - The grid - The story of the electricity grid
• Still Burning Podcast by Kent Beck - Honest conversations about fear, uncertainty, and what it means to build things when the ground keeps shifting.
• The Agentic Review - Closing the gap: AI Agent Hype and Production Reality - an interesting conversation discussing context engineering and deterministic memory systems that scale.

Posts

• Martin Fowlers Fragments: April 29

  • Chris Parsons has updated his guide on using AI to code https://www.chrismdp.com/coding-with-ai/
  • and much more

• Claude Code, Cursor and Codex hacks

• Hey, N00b, We Didn’t Hire You to Complete Tasks - Article from Kent Beck: Some of y’all (we’ll call them A’s) will be amazing game-changers, making everyone around them wildly more productive. Many of you (B’s) will be solid performers. Some of you (the C’s) won’t be here in a year.

• Interrogatory LLM - From Martin Fowler on using an LLM to interview a human in order to build context for a new feature.

• 40 Essential White Papers That Will Make You Great at System Design & Architecture

• How to debug a team that isn’t working: the Waterline Model

• Why Senior Engineers Fail System Design Interviews

• Make the Domain Explicit: From Procedural Mess to Local Reasoning - a great article on taking code that works but has high cognitive load, and refactoring it into a much simpler state which helps humans and AI understand it and do a better job of reviewing it.

  AI models do not “understand” code the way humans do. They infer meaning, and depend heavily on what the code communicates through names, boundaries, and structure. This is also why long procedural methods make coding life harder for agents. It’s not necessarily length per se, but rather that the longer the method, the more likely that it mixes multiple actions and responsibilities into one weakly described unit. That will confuse any agent.

• Technical Interviews Reject the Wrong Engineers - 20 years of observation, 50 years of research, and a framework for measuring the interview instead of the candidate

• The PRD is the highest-leverage artifact in software

  Large language models are not deterministic. The output is statistical. You cannot guarantee a correct answer. The only lever you have is the clarity of the context you feed in, and the relationship is direct: cleaner context produces better output, fuzzier context produces plausible nonsense that compiles, runs, and is quietly wrong. A good PRD is the best context you can hand an engineering agent

• Ktor + Kotlin Multiplatform: Full-Stack Type-Safe Development

• A fireside chat on the evolution of the developer craft Speakers: Addy Osmani, Richard Seroter, Aja Hammerly, Ciera Jaspan

• The Interactive Handbook on Data Structures and Algorithms This one-of-a-kind interactive book offers a hands-on approach to learning data structures and algorithms, featuring rich visualizations, code execution playback, an assortment of problems and an embedded Python environment. It’s designed to support both beginners and more experienced learners through active exploration and interactivity.

• I Tested Three Spec-Driven AI Tools—Here’s My Honest Take

  • A practical evaluation of modern spec-driven development tools, exploring how well they convert structured requirements into production code.

• From Vibe Coding to Spec-Driven Development

  • Tracing the shift in AI-assisted engineering away from casual, conversational prompting toward rigorous, specification-backed architectures.

• A Blast from the Past: SDD and the Software Factory

  • Adam Tornhill looks back at earlier software paradigms to contextualize Spec-Driven Development (SDD) and how vintage engineering concepts apply to modern AI agents.

• Fundamentals of AI System Architecture

  • A deep dive into core architectural building blocks required for designing scalable, production-grade AI systems.

Other

• Spec driven development course from JetBrains at DeepLearning.ai
• https://airskills.ai
  • Backup, publish, and sync your skills across your team.
• OpenScreen
  • OpenScreen: a free app for creating great video demos. Screen recording, auto zooms, annotations, effects - without watermark, commercially usable!
• Metapatterns GitHub Repository
  • A repository exploring structural abstractions and software design metapatterns.
• LinkedIn Text Formatter
  • A handy utility tool for applying bold, italic, and clean formatting styles to technical posts on LinkedIn.