Treating LLMs as sophisticated bullshit generators and designing systems that harness them effectively
Endless Method Definitions in Ruby
LLMs are coherence machines, not truth-checkers
Why you need to understand them in order to herd them
Why Narrower LLMs Outperform "God Models"
Code is both an asset and a liability
Part 2 on the industrialisation of software development
Part 1 on the industrialisation of software development
How GenAI alters the Not-Invented-Here dynamic
How restricted context windows help us build better agentic systems
Why metaphors matter when talking about AI
Moving fast doesn't need to break things
LLMs can't lie because they don't know the truth