Planck's Principle As Applied to AI
For people who follow my posts you will know I love a a good Paradox or Principle that can be used as an explainer with #AI.
Here is another that I think it particularly applicable when it dome to AI, Planck’s Principle.
Planck’s paraphrased observation is that “science advances one funeral at a time”, which means that new scientific theories often don’t win by convincing opponents, but rather by outlasting them as a younger generations are more open to adopting the new ideas.
As an example, when deep learning systems first demonstrated strong performance in reading medical images around 2016-2017, reactions were generally divided among radiologists.
Many senior radiologists were skeptical or dismissive, having spent decades honing interpretive skills that AI seemed to threaten. Some argued the technology would never match human judgment for complex cases.
Meanwhile, radiology residents and younger physicians were often quicker to see AI as a tool that could enhance their work, handling routine screenings while freeing them for complex interpretation and patient interaction.
A decade on, AI-assisted radiology has increasingly become standard practice, not because the skeptics were persuaded by arguments, but because the field naturally evolved as newer practitioners who grew up with the technology moved into positions of influence, and because the results became undeniable.
The demonstrates the Planck principle i.e. the transformation happened less through conversion and more through generational succession (combined with results).
I am sure we will see this play out again i.e. Today, senior developers often dismiss AI coding as producing mediocre code or undermining their craft, whereas Junior developers are embracing them without the same friction. As the latter group matures into senior roles, AI-assisted development will not be a debate - it will be just how code gets written.
Prior AI Paradoxes and Principles:
Hey, great read as always, its fascinating how this principle really captures the human element of AI adoption. I wonder if the speed of this 'funeral' is accelerating with how fast AI develops, or if human nature will always dictate the pace?