Saturday, November 05, 2022
Back in the 1940's a fund was allocated to a group of 10 intellectuals who were into automata, probability and mathematical fields. A general notion amongst the general people at that time was that a machine cannot perform a task say X. So the scientists made prototypes to show that a machine can perform task X with better efficiency. But the approach used was exhaustive and because of the limitations of computational power, nothing significant could be achieved. Hence came the AI winter with people criticising the methods as a wastage of resources.
It was in the 1980's when Japan came up with a supercomputer and the West followed it, this again brought up the discussions among the AI enthusiasts. But because of the cost factor and lack of efficient algorithms we witnessed another AI winter.
With the development of genetic algorithms and neural networks in the 1990's we witnessed a huge AI BOOM which never looked back till date.
There are three forms of Superintelligence:
1. Speed Superintelligence
2.Collective Superintelligence
3.Quality Superintelligence.
Speed Superintelligence is a system which can do everything that a human does but at a much faster speed. And 'much' here means a million times faster. This is the simplest form of Superintelligence. Collective superintelligence is where a group of people with the same intellect come together to pull off a big task. This is difficult as a superintelligence but as humans we are only familiar with this because we see it in the forms of firms, organisations etc. Quality superintelligence is a system which can perform at least what humans can do but with a greater quality. For example, every animal has a brain and their brain is enough to sustain them in their environment, but compared to humans, it's weak. The difference in the quality of a human brain over an ape will be the same as the difference in quality of a superintelligent machine and a human mind. When collective Superintelligence is enhanced by improving the communication and cognitive linguistic complexities then we get quality superintelligence.
Each of the three superintelligences are capable of achieving the other two once we achieve any one of them. Quantity is a poor substitute for quality. Hence all the three Quality superintelligences dominate.
There will be problems that could be solved only by quality superintelligence and not the other two. For example, take a literary work. Will 2000 people together be able to pull off a work done by a single highly skilled artisan? IMPOSSIBLE
Digital brain >> Biological brain.
Our brain works at 200Hz and hence cannot perform long sequential tasks and perhaps does parallel processing. A microprocessor works at 2 Ghz. Moreover complex algorithms cannot do parallel processing. It would be awesome if the brain could combine these two things. Human brain stores a billion bits of data in its lifetime which is less than the storage of smartphones.
Speed(Neurons) ~ 120m/s
Speed(electric signals) is equal to the speed of light.
There are machines that are already better than humans in several domains. For example a Chess engine or in a scrabble game. But the irony is once a system is developed it is not considered an AI. It becomes a software. If blindfolded evolution led to Human Intelligence then a known evolution can definitely lead machine intelligence to Superintelligence.
WHOLE BRAIN EMULATION
Whole Brain emulation is a way to create an intelligent software by scanning and modelling your biological Brain. In Layman terms, we can say that you are uploading the information in your brain and storing it on a machine.
Check out this video to learn more and read my twitter thread to learn how the entire process of Whole Brain Emulation works.
See you next Saturday, have a great weekend:)
Cheers!