Some thoughts after reading this great blog post from Columbia’s Stats blog:

Where we are today, Large Language Models (LLMs) have become skilled at mimicking human intuition— taking into consideration all of the individual’s (well, the LLM’s scraping of the internet, in this case) experiences and perceptions and providing heuristic responses to prompts.

What’s really exiting is as we move out of the nascent phase of LLM AI we can now attempt to recreate the higher level of human cognition: reasoning. The question now is which LLM AI product will be the one to realize that a large investment in bringing in psychologists will help them reach that goal first?