News
When we look at current AI Models, we see a very interesting and common event occur between them. It seems as if during an AI Model's training it improves linearly for a majority of the training and then suddenly "collapses" into a much simpler and more accurate model (Watch THIS video if you don't know what this is referring to)
What if we could replicate this phenomena with certainty? What new possibilities would open up for us?
The goal of this publication is to answer that question, not with another question, but with a design. A design, that can be made today, with real, currently existing technologies. This is called 'The Sandbox'. (CC BY 4.0, APL 2.0)
And to propose theories of the universe, not as factual descriptions, but as hypothesis seeking to be proven or disproven by scientific measure and study. They have predictions that can be measured first, and only then should they be taken seriously in any way. 'The Sandbox' should be able to help simulate and test these as well, pushing the boundaries of modern physics. (CC BY 4.0)