AI and Tools
New tools deserve evaluation, not automatic adoption or automatic rejection.
Article
Series
Series collect related notes without forcing every idea into a heavy taxonomy.
New tools deserve evaluation, not automatic adoption or automatic rejection.
Systems last longer when change is expected and visible instead of hidden behind optimism and wishful thinking.
Notes on using software and AI to study attention, memory, reasoning, and learning.
A lab note on using a source generator to turn a small enum specification into a richer enumeration type.
A field note on using code coverage to reveal tangled responsibilities, hidden coupling, and the real cost of a big ball of mud.
The first answer is often a reflection of the current framing, not the real design problem.