What happens when I get bored… well… I get ideas… big ideas. Unreasonable ideas. Such is one idea: Let’s build a framework that might be, at some point, labelled as sentient.
We all know that all of the big tech companies like OpenAI are trying to achieve AGI, but what exactly are they trying to achieve? Honestly… a glorified Library of Babel that uses a near real-time Dewey Decimal System. Sure, great, but… so what?
To build an SAI, even if its narrow in scope, it needs to have Introspection, Reasoning, Reflection, Curiosity, Ethics and Synthetic Emotions - along with all of the other good stuff that makes AI an AI. Not just those, but it also needs a way to interact with the world through various devices.
And so… while this is a tremendous work in progress, it has been started:
While there might not be an update for a bit, I’m still working on it. One of the key aspects is the implementation of what I’m calling MetaQ*, which is a combination of Meta-Learning, Q-Learning and A* pathfinding. It’s slowly coming together and will eventually put up on github as a single package as well. Once I get that done, then we’ll get back to the actual SAI project and start putting together the feature extractors.
If, and this is a huge IF, I actually complete this project (it’s not like I’m being paid for it) and IF it actually works, well, hopefully it learns enough to actually start improving itself through Self-Optimization and Improvement. Who knows, we’ll see. At least I won’t need funding that would amount of the next 10 years of the entire global GDP to complete it - yes… I have a project that has that as an estimated cost to deploy it.
I guess I should also mention how I hope to achieve it and what kind of neural network(s) I’ll be using? Sure…
- Quantum Neural Network: I’m convinced this is needed, see → Quantum Biology
- Deep Neural Network
- Recurrent Neural Network
- Probably some type of multimodal Transformer - research is still ongoing
Anyways, yay? Back to coding…