Blog
Stanford's CS146S: The Modern Software Developer
An overview of Stanford's forward-looking course on AI-assisted software development and its core principles.
On Leading a Room Full of Experts
A summary of Ibrahim Diallo's article on the role of a leader in a team of specialists, focusing on translation, goal-orientation, and context-setting.
Scaling Python Data Science: Lessons from Coiled
A summary of a Talk Python podcast episode discussing the trade-offs of scaling data science workflows, focusing on Coiled's approach to environment replication and productionisation.
Timeless Leadership Principles from Jocko Podcast 251
An exploration of foundational leadership principles from General Bruce C. Clarke's 'Guidelines for the Leader and Commander', as discussed on the Jocko Podcast.
Understanding the Five Doors of Influence
A framework from Andi Roberts for influencing others by understanding their preferred communication style, from logic to inspiration.
From Code Generation to AI-Native: Insights from the Front Lines
An analysis of a panel discussion with Anthropic, Trae, and Replit, connecting their best practices for AI-native development to the evolving role of the software engineer.
Revisiting Vibe Coding: An Analysis of AWS's Practical Guide
An analysis of the AWS guide on vibe coding, comparing its structured, tool-centric approach with earlier conclusions on agentic workflows and human-led architecture.
You Should Delete Tests
A summary of André Arko's argument that tests should be deleted when they no longer increase confidence, with a focus on flaky, high-maintenance, or irrelevant tests.
A Practical Guide to Coding with LLMs: Dos and Don'ts
A structured approach to using AI coding assistants, moving from contemplative prompting to scalable agentic workflows.
Is AI the Next Microprocessor or the Next Shipping Container?
An analysis of whether AI will create new wealth like the microprocessor or primarily benefit incumbents and customers like shipping containerisation.
Working with AI Wizards and the Need for Explainability
Ethan Mollick's 'wizard' analogy for AI highlights a growing problem: as models become more powerful, their processes become more opaque, creating a critical need for explainability.
Contrasting Memory Philosophies: Claude's Explicit Tools vs. ChatGPT's Automatic Profiles
A look at the different approaches to memory in Claude and ChatGPT, highlighting the trade-offs between transparency and convenience.
A Structured Methodology for Disciplined AI Software Development
A summary of the 'Disciplined AI Software Development' methodology, which provides a framework for managing AI collaboration in coding projects.
The Tension Between Legibility and Efficiency in Software Companies
Exploring the concept of legibility and how large organisations often prioritise it over real efficiency, based on an article by Sean Goedecke.
Frameworks as Mental Handrails for Navigating Complexity
A summary of Mike Fisher's article on frameworks as cognitive tools, with reflections on their application in TOGAF and Systems Engineering.
Where's the Shovelware? The AI Coding Claims Conundrum
An analysis of why AI coding tools haven't led to a surge in software production, despite widespread adoption and bold productivity claims.