There is a classic management principle called Parkinson’s Law, coined by historian C. Northcote Parkinson in 1955:
“Work expands so as to fill the time available for its completion.”
In simple terms:
Give a project six months, it will take six months.
Give it two weeks, it might actually ship in two weeks.
We’re seeing this play out in the AI race.
Some companies are shipping early and learning in public. Others are waiting for perfection before release.
Take the recent delays in advanced AI assistants from major tech companies. Rather than launching imperfect systems and iterating, some firms are holding back until reliability meets internal standards.
Both strategies have merit:
Ship early
• Faster feedback
• Faster learning
• Faster market adaptation
Wait for perfection
• Higher initial quality
• Lower reputational risk
But Parkinson’s Law reminds us of something important:
When timelines expand, complexity expands with them.
Scope grows. Committees grow. Momentum slows.
In digital transformation and AI adoption, the winners may not be those with the most resources, but those with the shortest learning cycles.
Constraints create progress.
Iteration creates learning.
Learning creates advantage.
So the real leadership question is:
Are we giving innovation enough time to learn — or so much time that it stops moving?
