Hi just wanted to put what I think are some of the major ingredients for AGI together I’ve made posts on these topics before now I would like to put them together in one post I’ve been looking into any major AI advancements and potential solutions that I can find and here’s my opinion on what things we might combine in order to create the worlds first AGI going to be quoting myself from previous posts I’ve made
Steps towards AGI in my opinion
The AI hardware problem
In my opinion this could potentially be solved by some of the ideas stated in these 2 videos
https://youtu.be/owe9cPEdm7k “titled the AI hardware problem posted by the channel New Mind it talks about the potential of older analog computer tech as a less power hungry way for machine learning type technology”
https://youtu.be/19fjsk9blB4 “this is a video titled Neuromorphic: BRAINLIKE Computers by the channel CoreTeks which i think elaborates on this idea a little more including potential solutions to the von neumann bottleneck” also we might want to limit the processing speed of AGI in a human like body to interact with or else it might be like how the flash experiences things where everything happens really really slow to it might make for problematic conversations between humans/AGI just a side thought also I’d say the biggest takeaway for hardware is combining cpu with memory to potentially overcome the von neumann bottleneck alongside combining both digital and analog computing together so the we might make more power efficient and eventually more powerful overall AGI hardware in one of these videos they use termites as example as to how neurons might work saying they dont know what they are doing yet they do it anyways without a “central command” so to speak much like how neurons function so R&D in that may help as well
Now for the AGI software problem and while I don’t have much for this one out of all the algorithms for AI that I’ve seen while looking into advancements in AI I think that advancing OpenCog, AI-DSL, and some more R&D on this https://blog.singularitynet.io/an-understandable-language-processing-3848f7560271 and Agent57 https://deepmind.com/blog/article/Agent57-Outperforming-the-human-Atari-benchmark and finding out what Agent 57 does right and wrong and working from there alongside the most advance AI algorithms that have been developed by everyone else is at least a small part of the greater whole that is AGI in my opinion sorry I don’t have the answers to these questions i just think that solving these problems will help lead to AGI all credit goes to rightful owners
And if you agree with me please help spread this information to the right people so that AGI can be beneficial to ALL life not just the 1 percent of the 1 percent thank you for taking your time to help validate and or invalidate my opinion on what I consider to be the major ingredients to AGI soup I bid you farewell have a wonderful day/night.