What is it thats makes the human mind unique?
One hypothesis I’ve been building up is that the mind has a limitlessly inspectable simulation engine. Think of a video game engine like Grand Theft Auto, where you can walk around and explore a world, only infinitely.
From the first day of life, all organisms are building a model of their world to survive and thrive, but the human’s model is unique in that it can store more information and introspect that information in many different ways.
- We build a physics model of our world from day one. Human babies are interacting with the world and grasp intuitive physics: how objects feel, friction between the body and floor, objects fall to the ground, objects can be stacked and fall over, liquids are not 100% solid, objects are solid and don’t just disappear, different objects have different weights, ball shaped objects roll around, hidden objects don’t cease to exist, and more. This physics engine is the base of all aspects of our lives and gives meaning to words like “rough” and “smooth”. Many scientific papers have been written about the internal physics models of babies.
- We can understand cause and effect of almost anything by playing scenarios in our heads and thinking about potential outcomes. This cause and effect engine seems to run at both a conscious and subconscious level. At the subconscious level, you can see this happen when driving or doing extreme sports, we make quick decisions to avoid harm. The subconscious system seems to work best with alerting you to potentially dangerous situations. At the conscious level, humans are constantly thinking about the consequences of potential choices. If I drink coffee at this time of day, it may effect my sleep. If the trees around the house are on fire, the house will probably also burn down. If I take this job, my daily commute time will increase by 2 hours, pass. We can change out objects and concepts like variables at any level of the engine and observe the effects. This engine is like our physics engine, but more abstract and with a longer time horizon. I suspect this cause and effect engine is just the next level up of our internal physics engine.
- Extreme goal optimization. All organisms are constantly trying to achieve a particular goal, whether it is eating, avoiding predators, or resting. The human mind takes this to an extreme level. We can change our goals to be anything specific or abstract. Someone can decide they want to be famous and achieve that without knowing the specific details at that time. We can also focus on specific goals like while playing video games: kill all the enemies, or the goal could change instantly to live as long as possible, or the goal could change to kill only specific enemies. The goal can keep changing and we could still be able to progress. Contrast that with computers, we cannot build computer systems that can change or work with multiple goals. Each goal must be painstakingly built for a computer and any modifications to a computer’s goal will typically result in failure. A human’s goal can be long term, days or years out into the future, where as all other organisms’ goals are real time.
- A time traveling engine. We can play any scenario we want in our heads and see potential outcomes in the future and past. The planet will probably continue to become hotter due to human activity. If World War II didn’t happen, we probably would have had less deaths in the 1940’s. We can think in short term horizons or long term horizons. This is similar to the cause and effect engine we have and may be the same thing, but the focus on time travel is important because we can focus on the past to potentially modify our current plans.
- Observe things from different points of view. This is similar to a video game engine where you move to a different scene or camera angle. Humans can easily switch out the point of view from themselves to another perspective. This ability allows us to understand the concepts of I, you, them, and the world. We can use this ability to understand other people’s mental models. For example, as someone who needs to drive to be able to work, you may think that the right to drive is important. If you know someone who lives in a “bicycles only, no cars allowed” city, you may be able to understand why they think cars are unnecessary to live a good life. We can emphasize for others because we can see their life from their point of view. We use this ability to make decisions, especially when multiple parties are involved, think governments.
- Models on top of models. Models and simulations are basically the same thing, but I like to think of models as simpler simulations with less moving parts. Besides the models I’ve shown above, the human mind can create new mental models and then use them at any time. We have a model of 24 hour time cycles, analog clocks, driving, day/night cycles, biology models, birth/death models, basketball models, writing blog posts models, growing crops models, virus models, having a family model, etc. We constantly create new models such as mental models, abstract models, physical models, and probabilistic models. And then take those models and build other models on top of them. The human mind may be composed of models all the way down.
- Focus/ignore engine – The ability to focus on aspects you want and ignore everything else allows us to focus on specific information. Animals use this ability to avoid prey. Of course with humans it is more powerful. This ability allows us to amplify a specific thought to produce more related thoughts and expand the idea. We can change our focus in the short term and long term. For example: you are skiing , there is snow, wind sounds, pain in your legs, other people, rocks, birds flying, lots of sensory information coming in, but you are able to focus on the act of skiing to get you down the hill safely. Another example, you may recently learn about the difference between vans and minivans where as before you thought they were all the same, and now you notice both minivans and vans going forward. A human’s focus engine can change and adapt to different situations. Compared to computers, they have primitive focus engines that need to be hand programmed. Focus is usually called attention in artificial intelligence research and is an active area of research. This focus ability can sometimes lead to negative effects where a human focuses too much on certain stimulus and it creates addiction to things like gambling, greed, alcohol, etc.
If you were to able to build out just one part of the human simulation engine mentioned above into software, you have most likely solved true artificial intelligence, good luck 🙂
ai could tranfers what happen to something under certain circumstances to a avatar in a simulation.you could do this with a type of gan.
the ai could be programmed to avoid situation if avatar is distorted avoid situation.
A time traveling engine.
We can play any scenario we want in our heads and see potential outcomes in the future and past.
The planet will probably continue to become hotter due to human activity.
If World War II didn’t happen, we probably would have had less deaths in the 1940’s.
We can think in short term horizons or long term horizons.
This is similar to the cause and effect engine we have and may be the same thing,
but the focus on time travel is important because we can focus on the past to potentially modify our current plans.
————————————————————————————————————————
human notice what has been done and not done.
ai will need to think about what is thinking.
actions and objects can be seen as tool to accomplish a goal by the ai.
have you seen this.
Replay in Deep Learning: Current Approaches and Missing Biological Elements
https://deepai.org/publication/replay-in-deep-learning-current-approaches-and-missing-biological-elements
Observe things from different points of view.
This is similar to a video game engine where you move to a different scene or camera angle.
Humans can easily switch out the point of view from themselves to another perspective.
This ability allows us to understand the concepts of I, you, them, and the world.
We can use this ability to understand other people’s mental models.
For example, as someone who needs to drive to be able to work,
you may think that the right to drive is important.
If you know someone who lives in a “bicycles only, no cars allowed” city,
you may be able to understand why they think cars are unnecessary to live a good life.
We can emphasize for others because we can see their life from their point of view.
We use this ability to make decisions, especially when multiple parties are involved, think governments.
—————————————————————————————————————————————-
example of observing things points of view.
the ai could have an input device in it’s brain to register when it has a thought something.
the ai could have input device out side it’s brain to register inputs from outside itself.
it could compare those inputs to see how similar they are.
if they are similar enough it could think about what that being is going to do next.
i think gpt 2 and 3 already does this.
instead of predicting the next words it could predict the next action.
combine that with reinforcement learning and you would have a great ai system.