The Myth of Artificial Intelligence: Why Computers Can’t Think the Way We Do - Erik J. Larson

This post has 17 comments with rating of 5/517

• 
  caesar963
  October 23rd, 2021
  

  

  I abduce that those metal muthas are doin’ their best, dammit!

  The Myth of Artificial Intelligence: Why Computers Can't Think the Way We Do - Erik J. Larson
• 
  Thamus
  October 23rd, 2021
  

  Machines will only be able to pass the Turing test because humans are, by their nature, all too willing to meet machines half way by dumbing themselves down.

  The Myth of Artificial Intelligence: Why Computers Can't Think the Way We Do - Erik J. Larson
• 
  howlafist
  October 23rd, 2021
  

  > But humans don’t correlate data sets: we make conjectures informed by context and experience

  And the difference here is …what?

  There are 4 things needed to converge current methods with AGI:

  - live updates
  - global model coherence
  - attention and motivational modules to handle and decide interrupts
  - guys like this (or their marketing teams) to stop misleading people about what a computer is

  The Myth of Artificial Intelligence: Why Computers Can't Think the Way We Do - Erik J. Larson
• 
  mklangelo
  October 23rd, 2021
  

  @caesar963 I suppose if you use words like, “abduce” that 99.9% of people have never known, you might sound smart. lol

  The Myth of Artificial Intelligence: Why Computers Can't Think the Way We Do - Erik J. Larson
• 
  abid
  October 23rd, 2021
  

  Makes sense. Machines could never create a newer model of Terminator. What kind of program could one write to make a robot able to innovate on such a grand scale?

  The Myth of Artificial Intelligence: Why Computers Can't Think the Way We Do - Erik J. Larson
• 
  pebbles188
  October 23rd, 2021
  

  The language of science is maths.
  All maths must be capable of running on a Turing Complete computer- in other words no randomness, just state n -> state n+1 determinism.

  It is trivial to prove that the Human mind (and the minds of all other ‘living’ entities) are a result of consciousness- a non-deterministic semantic pattern processing concept.

  Computer programs are simply maths in motion- syntax and not semantics. The Universe outside of mind is clockwork. Mind is not clockwork. Why this is and what this means cannot be explained this side of ‘death’.

  The clockwork Universe does not need mind to function- and therefore mind is an alien visitor/invader of the clockwork universe. Mind changes the Universe by allowing events otherwise statistically impossible as a result of energy probability equations. In other words free will allows the living to create outcomes that defy science- though the vast majority of the living are too ill-educated to understand this.

  In the 1950’s, some computer scientists (really computer engineers/mathematicians) believed in AI, and knew that long before our level of processing power (2021), artifical life (in code form) would clearly reach the level of the smartest life forms below Man. Why? Because the processing power of the brain and nervous system of animals was calculated- and thus when Man made computers reached this level, AI must surely rival animals. This never happened.

  Today AI is simply a term for a branch of Maths and algorithm design, and has NOTHING to do with the conscious mind of life.

  The Myth of Artificial Intelligence: Why Computers Can't Think the Way We Do - Erik J. Larson
• 
  caesar963
  October 24th, 2021
  

  @mklangelo - Er, “abduction” is right there - in the description! And is actually central to it; that’s the terrible irony here. You didn’t read/understand the description? When you ass-ert that 99.9% of people don’t know stuff, you’re really just talking abt yerself again, ain’t ya, pardner? And you’re on a book site - for some reason! Rofl! Don’t be bitter, cheer up, ffs.

  The Myth of Artificial Intelligence: Why Computers Can't Think the Way We Do - Erik J. Larson
• 
  howlafist
  October 24th, 2021
  

  @小石
  This is just wrong? The takeaway from gödel / turing’s most famous works is that much of maths *cannot* run on a turing machine, meaning it cannot be implemented at all. That mathematics had couched in magical (impossible) eternalisms and infinities due to the religious influences of pythagoras, plato, aquinas, et al.

  And non-deterministic turing machines exist, and can resolve the exact same set of languages as can deterministic turing machines. All nondeterminism means in practice is the taking in of arbitrary / unknown bits from ones environment. Terms like deterministic, non-deterministic, probability, and randomness are only defined with respect to, and are measures of, what one personally knows and doesn’t know.

  Free will is a feeling you get because you, the attentional agent bit of your brain, are not all-knowing and do not make up the entire brain. So there are decisions made by the other parts of the brain that seem novel (undetermined) when you notice them but also have a feeling of belonging to you.

  That the human mind is a universal computer is easily proven: it can emulate other universal computers. The only specialness-claim left for someone to make is that it might be a hypercomputer, i.e. something impossible to implement and that would necessarily give us magical predictive powers we don’t seem to possess.

  The Myth of Artificial Intelligence: Why Computers Can't Think the Way We Do - Erik J. Larson
• 
  howlafist
  October 24th, 2021
  

  (Or nondeterminism has in computability theory the meaning of “try every path simultaneously and take the one that works”; either way)

  The Myth of Artificial Intelligence: Why Computers Can't Think the Way We Do - Erik J. Larson
• 
  ssafe05
  October 24th, 2021
  

  I tried to download this, but could not work out which of the pictures were of crosswalks. Darn!

  The Myth of Artificial Intelligence: Why Computers Can't Think the Way We Do - Erik J. Larson
• 
  nugaseya
  October 24th, 2021
  

  “The question of can a machine think is about as relevant as the question of whether submarines can swim.” -Edsger Dijkstra

  The Myth of Artificial Intelligence: Why Computers Can't Think the Way We Do - Erik J. Larson
• 
  caesar963
  October 25th, 2021
  

  “Edsger Dijkstra” - imagine the scrabble score on that name. Beyond the dreams of avariciousness!

  The Myth of Artificial Intelligence: Why Computers Can't Think the Way We Do - Erik J. Larson
• 
  Kalamarironman
  October 31st, 2021
  

  

  Important read.

  THANKS!

  The Myth of Artificial Intelligence: Why Computers Can't Think the Way We Do - Erik J. Larson
• 
  halatukit
  January 11th, 2022
  

  Machine learning is real. However, anyone who is honest will tell you it barely works. Everyone else dying to work in a hot field will drool over the hype and coolness factor and delude themselves. A team of Python programmers backed by google money, working 7 days a week can make a machine play Go. Thats not intelligence. Though they bill it as such. “Artificial Intelligence” is a marketing term, bloated with hype, no different from Virtual Reality, or Quantum Computer. All of these are buzzwords pushed by inept journalists and hubris.

  The Myth of Artificial Intelligence: Why Computers Can't Think the Way We Do - Erik J. Larson
• 
  cxsclutj
  March 13th, 2022
  

  Classical Machine Learning algorithms have efficiency and scalability issues they has nothing to do with thinking. On the other hand methods based on Deep Neural Nets has potential to solve problems that classical ML methods like SVM,Genetic Algorithms,Decision Trees etc. can’t.

  But still Neural nets are basically multiplication and summation of numbers. You can put infinite amount of data. It can produce models with 100% accuracy but that’s it. Let’s think about an robot with human shape that can detect objects,listen and reply, measure distances, predict weather, and when its battery becomes low it goes to nears charging what is missing? Well Let’s design a biological costume for it so from outside it is indistinguishable from one another human so it safely passes Turing test. But to make it indistinguishable we don’t need to AI. From a distance a well designed mannequin with dress will not be distinguishable either. There is a huge missing part.

  Instead of waiting centuries for machines to be like humans it is easier to analyze humans to find answer to the questions. are humans programmed? How can a software knows if it is a software or not? How can be coded? Since codes are just a sequence of 0s and 1s, and 0 doesn’t know if it is a 0 how can a collection of 0s and 1s know? From this point of view it seems impossible to find an answer.

  The Myth of Artificial Intelligence: Why Computers Can't Think the Way We Do - Erik J. Larson
• 
  dchant
  May 25th, 2022
  

  

  Thank you kindly

  The Myth of Artificial Intelligence: Why Computers Can't Think the Way We Do - Erik J. Larson
• 
  marbles300
  August 27th, 2022
  

  @pebbles188

  I really enjoyed your take but I think you’re way off the mark. Read Thinking Fast & Slow by Kahneman. Read Influence by Cialdini. You’ll see that our minds do work like clockwork, it’s just that the probability distributions of our functions are highly individualized and subject to internal, invisible stimulus of externally unknowable origin. For example, I don’t know how many people in these comments were abused or neglected as children but that stimulus provides an ongoing behavioral motivation that tips the scales on certain patterns and against others. We are not separate and apart from the world outside of consciousness. We are just the latest manifestation of universal rules in motion. We are the universe observing itself. Nothing more.

  The Myth of Artificial Intelligence: Why Computers Can't Think the Way We Do - Erik J. Larson

Add a comment (please log in before commenting)