Understanding the technology gap that will result from the advent of AGI

There's been quite a bit of talk about the artificial intelligence divide between companies and between countries. Some believe this technology gap will launch those at the cutting edge to new heights and decimate those who fall behind. What people are not talking about is the technology gap that will result from the advent of artificial general intelligence.

Most people think AGI isn't possible in our lifetime, but there are those of us who do. All it takes is the framework to make it possible (this is what I'm working on) and the resources to manually build out each knowledge domain until it reaches a point that the AGI is fluent and can continue to build and maintain a particular knowledge domain by itself. 

Mainstream AI practitioners think that manually building out each knowledge domain is too labor-intensive and impractical. They believe that knowledge can automatically be infused by processing text. But as many of us already know, processing text does not result in a human-level understanding of the world as we know it. We also know that you can't learn everything just by reading. A lot of human-level understanding is obtained by observing, experimenting, and doing. And think about all the complex formulas and tasks in science, technology, engineering, and math.

It's important to note that text-only trained answer bots that can statistically predict the correct answer to a question are not intelligent. Neither are text generators that statistically know how to assemble coherent sentences and paragraphs. These are merely clever mathematical ruses that may convince people they're intelligent. They have no human-level understanding of the text, context, questions, or answers.

Some insist that AGI requires the invention of a digital human brain with "billions" of neurons to rival our own. That it should be able to teach itself everything we know without human involvement. Even I don't think this is possible in our lifetime. Our tools and methods are currently too primitive to reverse engineer the firmware packaged in the brain at birth. Researchers are trying to figure it out, but we still have a long way to go. Given what we know about the brain to date, no one on the planet can precisely tell you how a person's brain processes, stores, recalls, visualizes, or recounts their favorite TV show or movie. Not the images. Not the dialog. Not the sequence of events. Conceptually perhaps, but not down to what happens in each brain cell. If we don't even understand how basic storytelling works, how will we ever duplicate (vice artificially replicate) the brain's more complex and challenging functions?

If text processing won't result in AGI and creating a digital brain isn't possible, how does one create AGI?

AGI can achieve an initial human-level understanding of each knowledge domain through a combination of specially written/generated documents, real-world simulation/experience, by directly being taught, or by someone manually entering the knowledge as humans understand it - to include the use of full-body motion capture bodysuits. AGI needs subject matter experts to help it make sense of experiences and information, answer questions, and validate that its understanding is correct. Except for manual entry, isn't this what we do now to educate ourselves and others? There are no mathematical shortcuts. These labor and time-intensive methods are the only way.

But who has the means to enlist teams of people to build out the vast amount of knowledge required to seed AGI? Who has the ability to ensure that the knowledge is represented consistently and uniformly across all forms and scopes of knowledge? Only large corporations and nation-states. And the list of entities with the wherewithal and willingness to undertake such an endeavor is very short.

The first entity to achieve AGI will be so far ahead of everyone else that it will be impossible for others to catch up. The AGI seeding process will also expose what's required for it to progress to superintelligence. Will the entity share its technology with others? Or will it keep it for itself and create business units to dominate all industries until it's the only one left standing?

If you want to learn about my AGI framework that makes AGI possible, follow me here on LinkedIn. I'm releasing additional information over the coming months in preparation for a product launch.