MIT 1700 Lecture Week 4 January 25, 2011
Alan Turing and the Turing Test
- According to him machines and humans can think the same, there’s no difference. Not a lot of
people were happy with this so he answers 9 objections that and prove that they can think
- Thought is the product of an immortal soul. Souls are given by god only to
humans. He answers: the prejudices of an unscientific past that relied on
scripture have been erroneous (believing the earth was flat, etc.) There.
- Head in Sand
- The day will never come that machines will surmount human intelligence. He
counters: This is a prejudice of wanting to preserve our ‘specialness’ thus a
belief in superiority of human thinking can not admit to the facts of computing.
We have enormous egos so we won’t believe that machines can do what we can
- computers are limited and thus not like us. He counters: we are also limited. No
one can answer every single question put to him or her.
- There’s still a lingering mystery as to what consciousness really is. He counters:
consciousness isn’t locatable as such. You can’t communicate consciousness.
- Computers can’t enjoy desserts or life, or make mistakes. The fact that we can
enjoy Ice Cream is what makes us human. He counters: This assumes that
computers in their current form will always be ugly and repetitive. Computers
may be able to enjoy things one day.
- Continuity of Nervous System
- Our neurons don’t behave like switches in a computer. Turing counters: The test
measures communication outputs, not the internal structure of computers
compared to human brains.
- Lady Lovelace
- Computers can’t do anything it isn’t programmed to do, so limited in response
capabilities. Turing counters: So are humans. We are limited in response
capabilities too. To say that a machine can not produce surprise because it’s
bound by rules is stupid. That shouldn’t set limits to surprise.
- Behavioural Informality
- Behaviour is bound by rules. Turing counters: the task is one of human
imitation, not one that tries to replicate human behaviour. Why would we
expect the computer responses to be uniform, shouldn’t they be as unique as
- eg. Jabberwacky
- Soviet Union launched Sputnik during the Cold War.
- ARPA launched by U.S to counter the Russians (?)
- Computers at this time (1950’s) operated by batch processing: give it a problem, spits out
results a long time later.
- Licklider: What if we could think alongside out computers in real time? He developed SAGE
(Semi Automatic Ground Environment System) which was complex network connected by
phone lines to combat centers running on new digital machines.
- Delivered real time info on what was going on.
- Manifesto: Man Computer Symbiosis MIT 1700 Lecture Week 4 January 25, 2011
- Our future will be run by the best minds and the best computers (best of both
- But what’s the use of a super powerful computer that only works for one person? Licklider said
the future was all about computers replacing us, as a nucleus of a new civilization.
- Libraries would be accessible anywhere and publication wouldn’t take so long. To
connect info without being too centralized or diversified in disciplines. This prediction
- In a traditional library there was too much info loss. Some disciplines have terminology
that are similar in each discipline, but if you only look in one section you won’t find a
really good article that might be in a different section.
- Some books were only published in Russian, French.
- Licklider wanted to link information. You may forget the name of the Author, but an
integrated future library would tell you the name of it. Libraries in 1960’s didn’t help
you out this much.
- He envisioned a true ARPAnet. Precursor to the Internet.
- Funded by US Military
- Innovations from ARPA include mouse, keyboard, personal computer, computer
- ARPAnet can facilitate synergy: working towards a common purpose or cause.
- By networking computers together we could share info with each other in real time.
- You can send/receive simultaneously.
- ARPAnet went online in 1969 in UCLA. First attempt crashed.
- We can potentially share info as a commmunity. Digital Renaissance would take awhile
to impact us all.
- Maybe there would come a day when we’d have to program AI.
-Minsky says we should make machines that think like children. Their interest was in
discovering how children learn.
- Intelligence is an interconnected number of millions of agents. You learn through
communication with other agents and environment.
- Computers crash. When a computer is given two contradicting instances of equal value
it gets stuck in a loop and then crashes. Binary Yes/No system crashes.
- When we come to a contradiction in our thought, do we crash? No. This tells us that we
don’t think in a Binary way. We can always get a 3rd party to help us out with a
decision. We can adapt to situations that are unforseen. When we can’t make a decision
we can appeal to someone else.
- A computer is self contained and can’t appeal to outside source. Solution? Network
computers so that if it ever gets stuck, it can appeal to another computer programmed
slightly differently. Millions of computers connected to one another, the sum total of
those computers would be like a brain, capable of making de