MYMike Younginblog.unplugit.fi·Sep 10 · 8 min readKnowledge must be usable beyond its sourceA strategy can make perfect sense in the room where it was written and still be almost useless everywhere else. The people in that room remember the discussion, the rejected alternatives and the diffi00
MYMike Younginblog.unplugit.fi·Sep 1 · 9 min readAlignment Needs Bounded OwnershipWhen organisations talk about alignment, the conversation often moves towards shared goals, shared terminology and common ways of working. That sounds reasonable. It can also lead to a harmful conclus00
MYMike Younginblog.unplugit.fi·Aug 18 · 9 min readSoftware development has always been an alignment problemThe phrase alignment problem is now strongly associated with AI safety. In that discussion, the concern is whether an AI system behaves consistently with human values and intent. That is not the subje30
MYMike Younginblog.unplugit.fi·Aug 12 · 8 min readAI-assisted development needs boundariesMost writing about AI-assisted development seems to settle into three positions. There is the productivity account, where an agent is judged by how quickly it turns a ticket into code, the prompting a42KS
MYMike Younginblog.unplugit.fi·Aug 6 · 9 min readInternal, External, and Contract Models Should Not Be the Same ThingDiscussions about model separation often end up being discussions about boilerplate. One object arrives at a service boundary, most of its fields are copied into another object, and a similar conversi01M
