Theory of Intelligence
The algorithm of intelligence is always pursuing to create meaning out of chaos. It is about different pieces of reality (information) together in a way that makes it purposeful and valuable.
To make reorganising information possible, the information of reality needs to be able to perceived/retreived. Humans do this elegantly by having senses and a brain with memory which always aims to approach reality with a model.
Intelligence is tightly integrated with the concept of action: In order to perceive reality, there needs to be change, momentum, processes.
It doesn't matter if these processes are automatic or require manual action. It is all done to perceive reality better. Also creating purpose out of chaos is sometimes done with automatic processes and sometimes with manual, conscious actions that are taken.
In order to grow/survive as an intelligence, it needs to sustain itself by finding a place in reality that is not hostile. For humans, for example, this means consistency in food and community, alignment of the climate of your habitat with what the human body requires, and opportunity for reproduction.
What is a business?
I think a business can also be seen as a form of an intelligent algorithm that goes beyond a single individual. It is a group of algorithms and life-forms that bring purpose and reality closer together by acting.
A good business needs to balance cost and reward. If there is more cost than reward, the business will fail, if there is more reward than cost, the business grows. A good business is always seeking to increase future reward relative to its cost. A good business sometimes invests (takes risk) by temporarily having a higher cost because it knows it will likely give a larger reward in the future.
A business is a form of intelligence.
In order for a business to survive and grow, it needs to sustain itself by creating a place in reality that is not hostile. It often does so by having a bigger influx of cash than outflux. It often does so by being of value to humans or other businesses.
A business learns by doing more of what works and less of what doesn't.
A recipe for artificial general intellgence?
Large Lanuage Models do achieve purpose, but it's still very niche.
A higher order algorithm that can retreive information and take action to increase purpose, is very similar to what humans do setting up a business.
If I can leverage the capacity of Large Language Models to work with chaotic information and turn it into meaning in a consistent way, I think we have something interesting.
Self-reflection of datastructures is important: since retreival of information has a cost, its value (and purpose) needs to be reflected upon in order to know if the value is higher than the cost.
What are the questions innate to us we ask ourself every day in order to fulfill our purpose to survive?
With ActionSchema I strive to create a framework for intelligence. Information that grows itself - from chaos to structure - with a given purpose.