As I read, it was not that I wanted fiction, but wanted more story about "why?, and where? and when?" Those are the execution of the "what? idea" of writing. Beginning with "I am no expert, but" is an excuse: as a reader, I don't care so much about expertise as "entertainment". Regardless of subject or intent, sing me a poem.
I think most people would probably not like your approach, and think of it as way too sci-fi and idealistic given the limitations of current tools. They would suggest studying AI a lot more before you attempt to really think of anything. Things that seem easy can be hard, things that seem hard can be easy. And it's not clear that we will ever really recognize superintelligences...if we program something, then we know how it works, and if we know how it works, we'll just say "Oh, it's an algorithm" and not ever appreciate the extent of machine capabilities.
I do like how you handled "consciousness" though by redefining it to mean 'examine your memory and then learn from it'; artificial intelligence is really just one big word game and if you can change your definitions properly to something achievable, then you might actually achieve them. Then again, you actually have to do it and that seems more difficult than talking about it. Most people will likely disagree with your "consciousness" definition and will prefer a more vaguer version that would be impossible to be built...
My suggestion is two-fold:
1) Learn AI first and find out its tools and limitations. Struggle through those boring AI lectures, or see if you can find any GitHub Machine Learning libraries so that you can start working. Also, try some side projects where you use AI to solve real-world practical problems, before you attempt to figure out how to induce the existence of an AGI. I'd even suggest dropping the existence of an AGI (there's no guarantee it can even exist anyway) and focus on a specific part of AGI: inducing emotion or consciousness by itself might be "pretty cool", and if you can do that without needing an AGI, and be able to apply them to real-world problems, then people might even accept those algorithms as being valid (though they may avoid the pseudo-scientific terms we're applying to them right now).
2) Look at the "Mario Lives" videos[1], where programmers were able to successfully represent emotional states and conversations in an AI that plays Mario. May not match your ideas about emotion and learning, but it is a sign that you can do anything you want so long as you change the definitions of words to be what's practical.
Thanks for your feedback Tariq.
To give you some context:
-I have taken AI as a course in the past (college 10 years ago), and I'm a computer engineer so I have the basics of ML,AI,etc. and I'm familiar with the libraries that are out there and the body of knowledge.
- I find that when you go deep into a particular topic or field and read lots of research you tend to become immersed and start to think like everyone else in the field. There is some wisdom to Zen's mind beginner's mind and I wanted to let that Beginner's mind explore freely and express its thoughts before my reimmersion in the research.
- I will definitely look at Mario Lives video thanks for pointing it out.
- I guess I was looking for someone to challenge the ideas as ideas more than challenge my credentials / approach but it's all fair.
- I'm not interested in solving a particular problem at this point in time in AI. I think everyone is taking this angle and everyone is trying to do this. I'm already in Tech, in a fairly good position, and solving real customer problems. I'm simply following my curiosity and a more grandeur goal and doing it just for the sake of following my curiosity.
Good luck.