0 Members and 1 Guest are viewing this topic.
This situation establishes an entropic potential or a potential based on the second law; entropy. The ion segregation causes a lowering of entropy and the need to increase the entropy.
Ions in water prefer to mingle into a uniform solution and not bunch up with their own kind.
The neuron deliberately creates a potential based on entropy.
There is a built in push at every dendrite, axon and synapse to increase entropy; increase complexity. Part of the entropic potential adds complexity via further branching and synaptic connections. The brain's support hardware and firmware is also designed to help regulate the flow of bulk entropy by giving it paths of least resistance. The neurons quickly reverse the blended ions locally, with ion pumping and exchange, resetting the entropic potential, for continuous learning potential; need to add more complexity.
To have an AI it is possible to program it or it is possible to recreate the brain but with the electronics parts, I prefer the hardware. It is a true artificial intelligence. The goal here it is not to program the brain but let the exterior program it, like life does the job with brains (not only human but animals too). With these method it is possible to have the power of the human brain, 1000 times faster than the human brain, power consumption of 2000 W and a size of a PS/1 (IBM) with the nowadays technology of the chips (except the current clamp inside the chip):
With these method it is possible to have the power of the human brain, 1000 times faster than the human brain, power consumption of 2000 W and a size of a PS/1 (IBM)
for continuous learning potential; need to add more complexity.
To have an AI it is possible to program it or it is possible to recreate the brain but with the electronics parts, I prefer the hardware. It is a true artificial intelligence. The goal here it is not to program the brain but let the exterior program it, like life does the job with brains (not only human but animals too). With these method it is possible to have the power of the human brain, 1000 times faster than the human brain, power consumption of 2000 W and a size of a PS/1 (IBM) with the nowadays technology of the chips (except the current clamp inside the chip):1/ The first problem is the to compute the N additions from the 10000 neurons because at each connection there is a weight. For that, the solution is to use a current clamp to add the currents of N wires. Each weight is conserved by transistors, I need some transistors to adjust the weight, and it is better to transport all the wires with their weight (2^n), a standard electric neuron in a animal/human uses 256 weights, like that there are less transistor in commutation, to increase the velocity of the brain. 2/ It is possible to read all the 100e9 neurons in the same time. For that, instead to use a transistor for each neuron it is possible to use a code (a combination) from around 120 isolated galvanic voltage sources. The transistors are used to connect the isolated galvanic voltage sources 2 by 2, it is not read the current from a dot A to a dot B but to test the connection of 2 sources (are they connected or not ?). It is easy to transport 120 wires (the electric voltage sources) inside a chip and it don't consume a lot of energy (the problem of heating). The combination is like AB & DE & etc. It is necessary to test with the transistors the connection between 2 voltages sources. To set ON a neuron when the value of the current clamp is reached: set ON the transistor of the combination of that neuron. To connect a neuron to another one: set ON the transistors to have the combination of that neuron.3/ The controller needs to adjust the weight of the neurons and choose what neuron is connected to another. Maybe the most difficult part to imagine. I would choose the solution to use 2 electronic brains to have one, build 2 electronic brains, one controls what neuron is connected (and their weight) the another electronic brain. That could help the next step (fonctionality of the emotions). Maybe more than 2 electronic brains to have one is better to increase the intelligence.4/ It is necessary to have an equivalent of the BIOS (for the PC) for the electronic brain I call it BIOSE something like the emotions. Maybe this little program can be initiated by itself (the spark of the conciousness), just connect sensors and actuators and let the neurons connected to themselves. Maybe it is possible to prevent ALL the programation of that BIOSE, just build enough electronic brains, and connect each electronic brain with sensors/actuators/internet/etc. Let live these electronic brains together to see if the spark is lightning on. 5/ To increase the counscious: cut the sensors or actuators or the links with the others electronic brains during a time. It is a new intelligent entity, not controllable with a program if the BIOSE start by itself, just the hardware configuration. It is also possible to copy easily the electronic brain to replicate it.