095 - Accuracy and Completeness

Accuracy and Completeness favor fewer steps, and shorter hierarchies.

Complexity Bottlenecks and Specialization favor more steps and deeper hierarchies.

After revisiting Thomas Malone's book "Superminds" and looking over another new framework today this particular trade-off came to mind.

Each additional step of inquiry produces some non-zero loss of information, whether by human cognitive bias or an algorithmic proxy. The act of reducing the scope of information and focusing on specific points is itself a subset of cognitive biases, imposed on that information at the very first step, and again at each step thereafter.

Humans have adapted to operate in small collectives, utilizing specialization, but under today's systems that inherently sets an upper bound on accuracy and completeness, based in part on the depth of the process. Gradual expansion of knowledge and technology also creates a drift, and fault lines between specializations form, until they reach critical thresholds that compel new specializations to fill those gaps. The gaps are constantly emerging, so some are always present, and information is constantly lost to them.

There is a solution to this problem.

The first step to overcoming this is deploying systems that overcome the Complexity versus Cognitive Bias trade-off. In doing so, a scalable entity is established that retains the maximum quality and full scope of information, on whatever domain and within whatever scale it is applied to. That information may also be iteratively updated and improved over time. This prevents any fundamental losses to accuracy and completeness.

The next step is to connect all tiers in any existing hierarchy to that maximum quality source. By making this direct connection any losses of information incurred at each step are minimized. Those losses of information may be further minimized through developing and applying individualized models of understanding for each individual who communicates with that maximum quality source. This allows for the more effective and individualized communication of any required information to that individual, as well as a better and iteratively improving understanding of that individual's responses.

There are more steps, but these two alone can and will revolutionize industries. The limiting factor remains finding that ever-elusive unicorn, the competent and/or ethical investor. Less than 0.5% of the funds being poured into the world's 2 most obvious Ponzi Schemes could bring the necessary technology to market within about a year. The added value would, in turn, be several orders of magnitude beyond anything "(de)Generative AI" has produced, while also remaining net-positive in their impact, unlike Degenerative AI.

How accurate and complete is your information?