113 - Complex Paths
While categorization is imperfect, and a form of cognitive bias, it can be useful in moderation. People active on LinkedIn can be categorized as "those who actively seek to learn", "those who passively learn, but only when emotions align", and "those who seek confirmation of a hardened viewpoint (non-learners)".
Interaction with the first group is reliably productive, and interaction with the second group can be productive if material is communicated in appealing ways. The third group is reliably unproductive, both in terms of interaction and the influence they exert on others, and they bring to mind the attached meme.
Some people learn actively, some passively, and some just run on an exercise wheel in a cage. It isn't the role of the educator to change if people learn but to educate those who choose to learn, to the degree that they choose to learn. Evolution didn't produce humanity because everyone chose to learn, but because those who chose to learn are the ones who survived.
I see examples of those who choose not to learn every day, some of whom make a living as Disinformation Brokers peddling AI snake oil, muddying the water with claims that are demonstrably false of LLMs and Reinforcement Learning (RL). There is rarely any point in engaging with such people and doing so comes at the cost of your own mental health.
The world has a sufficient number of people who choose to learn to keep one busy for a lifetime, and that number only grows when those who choose to learn are cultivated, as their successful cultivation pulls the fabric of society in the direction of education. The reverse is also true, making this look more like a "tug-of-war" on the fabric of society in practice today. The greatest mistake any educator can make is to assume that they have no opponent, as that reliably sends them flying into the mud.
Fortunately, this tug-of-war may become entirely one-sided in favor of education once more appropriate and capable technology is deployed. The other side has paperclip maximizers in the form of RL, which may be deployed more quickly, but which remain many orders of magnitude weaker and more fragile.