Wednesday, 17 January 2018

SUPERINTELLIGENCE BY NICK BOSTROM

’WORTH READING,  WE NEED TO BE SUPER CAREFUL WITH AI’’   ELON MUSK  CEO TESLA &SPACE X


It truly deserved to be read.Since with the beginning of humankind people have tried to obtain the status of superintelligence.In his book Nick Bostrom explain:paths,dangers,strategies of superintelligence.Nick Bostrom professor  University of Oxford,director,Strategic Artificial Intelligence.

We can’t give an exact definition of artificial intelligence,but we define it like any intellect that greatly exceeds the cognitive performance of humans in virtually all domain of interst.

What seemed to be unimaginable in the past ,today it ‘s happening artificial intelligence already out performs human intelligence in many domains.Few days ago at Stanford University reading comprehension test artificial intelligence machines built by Microsoft and Alibaba have beaten humans.Alibaba’s neural network scored 82.44 on test !!! and humans scored 82.304 and Microsoft 82.650

Early attempts to built artificial intelligence have been made for the first time by the mathematician I.J. GOOD who had served as statistician.In early 1980 Japan launched the first generation of Computer System Project that would serve as a platform for AI.In the 1990 new technique was launched which include neural networks and genetic algorithms (Good Old Fashioned Artificial Intelligence GOFAI)

Superintelligence have the following paths:
  • ·         Artifical Intelligence
  • ·         Whole brain emulation
  • ·         Biological cognition
  • ·         Brain computer interfaces
  • ·         Networks and organizations

I will not speak about those you will find more details in the book.Recently Elon Musk launched a company Neurallink Corporation which is working to link the human brain with a machine interfaces creating micron sized devices.Other path of superintelligence is whole brain emulation can been made by a detailed scan of the brain of particular human brain.This might involve stabilizing the brain post mortem through vitrification (a process that turn tissue into a kind of glass).A machine could dissect the tissue on thin slices which could be fed into another machine for scaning perhaps an array ofelectro microscopes.The data from scanners is fed to a computer for automated image processing to reconstruct the three dimension of neural network that implement cognition in the original brain.

The adoption of AI could also create not only good things but also bad things like huge unemployment a lot of jobs will disapeer ,capability control problem undesirable outcomes can happen.Introduction of machine superintelligence would create a substantial existential risk but will reduce many other existential risk.

No comments:

Post a Comment