Large language models have shifted my perception of how I see people. I now see each person as a language model in and of themselves.
These LLMS get trained on vast amounts of data, labeled data, text, images, video, etc.. With humans manually guiding the model whenever it misunderstand what it has been given. Their abilities and verbosity is directly correlated to the amount of energy and money spent by the companies that train these tools. There is a reason why the frontier models are produced by labs with mountains of capital. Capital to purchase data in bulk and hire the best AI engineers with million dollar salaries.
When I think back to capable people who have made an impression on me, it seems like their "training data" and resources that built them have been signficant. Usually born into stability, raised with some semblance of a quite life.. Obviously I don't know this for sure. The trend i've seen is that the highly logical people are also emotionally stable. That emotional stability laid the foundation for them to develop their critical thinking.