1 Comment
Jul 16, 2023Liked by Omar Shams

Interesting read. It seems that George Miller’s work which led to the concept of 7 +/- 2 as the limit of human working knowledge chunks (1 +/- 2 for too many people it seems some days...) needs to be considered in the discussion. I’ve often thought beyond the cost considerations correctly discussed in the article that much of org design and mgt work is an effort to get around first the physical limitations of one person’s work (e.g. exhaustion and hours in the day) and second the mental limitation of how many details a person can coordinate in planning (7 +/- 2; more or less.)

The article describes AI putting the correct information before a team. Given the AI “should” be able to identify who created all of the work product that went into the LLM within a company dataset, I consider it a possibility that organizational structure would mostly dissolve and an AI would simply assign team members to an issue based upon work product evaluations. A bit like how pitching is done in the major league. Never know when the game starts if the starter will have a complete game, middle relief, just relief, just a closer, or something else. Perhaps the most organic position in sports. Seems like AI could do the same with the vast majority of organization positions.

Expand full comment