Blog
Mastering Claude Code: Core Settings and Process for Developers
A concise guide to leveraging Claude Code's essential features for efficient and scalable developer workflows, focusing on core settings and process.
Functional Core, Imperative Shell: Separating Logic from Side Effects
Explore the 'Functional Core, Imperative Shell' pattern for building more testable and maintainable software by isolating pure logic from external interactions.
Securely Accessing OpenWebUI with Tailscale Funnel on Proxmox
A refined guide to setting up OpenWebUI in an LXC container on Proxmox, secured with Tailscale Funnel for public access.
Claude Skills: A Simpler Approach to Specialised AI Tasks
How Anthropic's Skills framework offers a more natural, library-like extension to LLM capabilities compared to MCP.
Agentic Engineering: Just Talk To It
Peter Steinberger shares his 'no-bs' approach to agentic engineering, focusing on direct interaction with models like GPT-5-Codex over complex tooling.
Digital Twins: From Industry to Ecology
Exploring the expanding application of digital twin technology, particularly its emergence in ecological research and wildlife management.
Avoiding Negativity Echo Chambers in Your Career
Marc Brooker advocates for steering clear of cynical communities to foster career growth and personal well-being.
The Critical Scrutiny of TOGAF: A Pragmatic Reflection
A balanced look at TOGAF's evolution, its claims as a methodology versus a toolkit, and the practical challenges of its application in enterprise architecture, reflecting on its continued utility.
TOGAF: A Pragmatic Starter Kit for Engineers
Why TOGAF still matters, how to launch it without drowning in paper, and the minimal habits that keep the architecture alive after day-one.
Vibe Engineering: The Senior Developer's Approach to LLMs
Exploring 'vibe engineering' as a structured, best-practice-driven method for senior developers to leverage LLMs, treating them as junior collaborators.
Extracting Knowledge from PDFs: Advanced Techniques and Local Solutions
Exploring methods for converting PDF content into structured, LLM-ready text, including insights from Andrew Ng on agentic approaches and a personal local tool.
Embracing the Parallel Coding Agent Lifestyle
Exploring how running multiple AI coding agents in parallel can enhance productivity for specific development tasks.
The Leader's Duty to Train and the Culture of Discipline
A leader’s primary duty is to ensure their people are trained and prepared for the worst-case scenario, fostering a culture of discipline and continuous learning.
Stop Avoiding Politics
A case for why engineers must engage with organisational politics, drawing parallels to Machiavelli's 'The Prince'.
Designing Safe and Effective Agentic Loops
Simon Willison's article on 'designing agentic loops' provides a practical framework for safely using powerful AI coding agents by sandboxing them in environments like GitHub Codespaces.
The AI Coding Trap
Chris Loy's framework for applying software engineering best practices to AI coding agents, treating them as lightning-fast junior engineers.