I came across these lecture videos as part of my on-going education. I was of course (unfortunately) unable to attend the lectures myself, so here are a few of my reservations about some of the "philosophical" statements. Almost as if I were sitting in the lecture hall :) 8:46 Opinions also have a truth condition, don't they? For the statement "I like apples", the natural language understanding system would have to be able to state that in order for that to be so, I'd need to like apples. 9:41 "Love is the most important thing in life" is a good example. Here, the opinion might be tied to personal ground truths. Humans are able to recognize this statement as such and conclude that for that to be so, love must be the most important thing in life for the other person. (Further, to assign a truth condition to the statement itself, either "enough" interactions could show that it is likely a generally true statement, or fundamental ground truths might be taught to the natural language understanding system, just as humans are taught norms and values growing up.) 10:28 An appropriate response of a natural language understanding system might be to learn a personal ground truth of the conversational partner first, and then to take it into account later in the interaction. 12:10 Memorizing personal ground truths and considering them when interacting with the conversational partner could be one of the many ways to empathize with a statement. "I am sad" is a personal ground truth that applies to the moment, because for that to be so, the other person must be currently sad. Understanding "I am sad" as a human often leads to softer conversation and more words intended to improve the conversational partner's mood - thus it might be argued that a system that understands natural language could be evaluated through its ability to empathize to show if it truly understands.

Human understanding is based on a few criteria: 1. If a trusted entity explains a statement is true, humans accept that statement as true without experimenting , examples of trusted entities are: Teacher, book, university , new paper , parents etc. This method eases us to extract abstract knowledge without understanding the complexity of truth. 2. But when we experiment with this abstract knowledge further, based on our outcomes, we correct our knowledge further. 3. Believe in God, somehow if we do not get any truth on a given statement and believe that God is taking care of the statement means it allows us to explore the abstract experiment further to find the truth. So humans invented God for this purpose.