Featured
•4 min read
You don't need your PC terminal to get the best from AI
Everyone keeps talking about having to learn Claude Code, but that won't be the way most people interact with AI in the future
G
General
Thoughts on startups, career development, and lessons from building tech products.
Read article
Featured
•11 min read
Agents work. Coding shows how.
How we can learn from coding agents when building AI systems for other parts of businesses.
•8 min read
'Human-in-the-loop' doesn't mean what most people think it means
How humans and agents should work together to create the most valuable and trustworthy outputs.
•6 min read
More information isn't always better information - Avoiding context rot in agents
Why cramming too much information into a prompt for an AI agent will give you worse results.
•11 min read
Six types of context your agents need - and where you can find each.
The 6 types of context for AI Agents in more detail and how to actually capture and make them usable
•6 min read
Your expertise is the product. Agents with context is how you scale it.
'Give it more context' has become one of those pieces of advice that sounds helpful but doesn't actually tell you anything useful or practical - here's what you should actually do.
•9 min read
Three reasons we're moving towards agentic automation
Why agentic automation is gaining ground: flexibility (update briefs, not code), compounding gains from better models, and systems that learn from feedback instead of fixed rules.
•7 min read
From flows to agents: the spectrum of automation
There are different types of automations, and the approach matters more than most people realise.
•7 min read
Lots of people are using AI, not many people are using it well
Most teams use AI; few get impact. Lessons from two years building agents on what actually works: connect real context, design agentic flows, and keep humans in the loop.
•4 min read
GenAI is Changing How Agencies Think About Automation
There are two universal truths agencies (and everyone else) have to accept in order to survive: Everything that can be automated should be, but not everything can be automated