Problem
AI lowers the friction of starting new work, and most people prefer starting tasks over finishing them. Working on many things at once can even look impressive.
What Goes Wrong
Many things start at once — branches, prototypes, experiments, tasks. You lose overview and get stressed. For the team it's worse: not only does everyone lose track of their own work, but what the team works on as a whole becomes overwhelming.
Consequences: attention fragments, and the team loses a shared picture of what matters. It becomes hard to explain to other teams what you need and what you can offer.
Symptoms: too many open loops, constant context switching, unfinished work.
Example
This is a very human challenge, but AI makes it worse — it lowers the bar for starting a task even further. It shows up most in teams focused on working hard rather than doing things right.
Solution
Limit work in progress. Finish, review, and learn before opening the next loop.
