Articles
Field perspective · Part 3
The most convincing thing about LLMs is also the most dangerous: they feel like they’re thinking. Decision-makers who confuse the illusion with the mechanism end up trusting the wrong things.
Understanding · Part 1
An LLM is a prediction engine that has learned to mimic the patterns of intelligent writing. There is no “person” in there — and being clear-eyed about that changes how you use it.
Field perspective · Part 2
A specific category of engineering investment — the elaborate techniques built to compensate for code’s rigidity — has stopped paying for itself. Identifying which parts of a stack still need it, and which parts were over-built for flexibility that never arrived, is now one of the more valuable things a team can do.
Field perspective · Part 1
Your company already records almost everything it says to itself. AI can now read all of it at once — and the organisations that move first will see themselves more clearly than their competitors can.
The Pragmatic Prompter · Part 1
Most AI-generated code looks plausible, compiles fine, and then slowly poisons your codebase with hallucinated assumptions. Three structural prompt changes fix that — before a single line gets written.
by ru5ty