Here's the biggest problem with trying to crowd-source research from beginners, you don't know what you don't know. You get 100k beginners and ask them to try to figure out AGI, they'll come up with a bunch of solutions that have already been tried thinking they're novel, but not realizing that it's been done before due to lack of experience.

I've tried to do something similar myself, not for AGI but for something else in another domain within machine learning. Thought I found gold and was a genius, only to discover I had just reinvented the wheel for an older technique that was abandoned due to not being feasible. As a result, all that happened was I learned firsthand and for myself why that technique was no longer used (and that it was ever used in the first place). It was a great learning experience, but that's all it was.

Depth of experience is invaluable. Research builds on past research, but in order to know what to build on and how to build on it, you gotta be experienced within the field. You gotta truly understand everything else that's already been tried.

I don't think you really appreciate everything that goes into research. It's a common fallacy for people who are beginners, like I said, you don't know what you don't know.

Of course, that shouldn't stop anyone from trying. You're more than welcome to take your advice. As for me, I am focusing more on novel ways to use advanced AI built by others in software applications. OpenAI just creates the tools, but someone's gotta use those tools to create something useful. That's where people with breadth of experience who lack the depth of experience necessary for rigorous research can excel. I'm not going to try and beat giant corporations with teams of PhDs and billions in funding at their own game.