Actually, neurosymbolic AI is an idea MIT/IBM seems to be very exicited about as the next breakthrough in AI. That what's missing from LLM's and neural networks is that they are missing the advantages of a layer of symbolic, old fashion AI reasoning added to it. We'll talk about this in class.

I have to admit I don't understand what the author is trying to say here by listing three (random) differences between us an AI. Is this list supposed to be a complete list? Is having "hopes" relevant to what the AI output 550 means?

Is this like a human banker explaining to Lucie why he assigned her a 550 by talking about which neurons firing in his brain led him to his decision or a human banker explaining to Lucie he assigned her a 550 because he went through schooling for how to evaluate credit scores and showing her his notes from class. (i.e. telling her about how he was trained)?

Are there other alternatives to understanding what the output could mean? Similarly, if in coming up with Lucie’s credit score I use a calculator to minus her negative score from a perfect credit score (800-250) and then the calculator displays 550 does this output represent her credit-worthiness or just what results when you subtract 250 from 800?

If I take a calculator, type in '800-250' then it will also output 550. But the way I got to '550' doesn't make it mean Lucie has a low credit score. I just picked a random number and then subtracted another number. It matters how I worked my way to that number that it means Lucie has a low credit score, so how did the SmartCredit work it's way to 550? (Do you think how the output is made is always relevant to what it means? Or how the output is recieved? Or both?)

Hi All! Our aim is to read all of Part I (so Chapters 1 & 2; up to page 50) to discuss in class on Thursday, Oct 2nd. I'll add annotations as I re-read the chapters this week, and you do too please!

Welcome! This is how you make notes. Reply under this note to say "hi!" and prove that you know how to use this app!