Ray Kurzweil The Singularity is Near

1
Ray KurzweilThe Singularity is Near

• Chapter 9
• Response to critics
• Presented by Dr. Chuck Selden

2
Responses are to criticisms ofthe Kurzweils
previous bookIn the Age of Spiritual Machines
3
The Malthusian criticism

• Exponential growth of computing power will cease
  as
• resources are depleted or
• excess heat melts the computer
• Response
• Computing efficiency will increase
• Smaller computing elements at molecular level
• Nanotubes and microtubules
• Massively parallel computing
• Reversible computing
• Limit will not be reached until computing power
  is trillions of trillions of times more
  powerful than human mind
• Challenge limit of control interfaces or
  input-output connections between micro- and nano-
  scale

4
Criticism from Software

• Exponential gains in hardware are not matched by
  evolution of software
• Slower doubling time
• Response
• New software techniques appear as breakthroughs
• AI techniques as an example
• Reverse engineering of the brains neural
  processes will yield efficiencies
• Increasing rates of understanding brain
• Architecture
• Operation
• Challenge Reverse engineering of the brain will
  yield hardware solutions, not software solutions

5
Criticism from Analog Processing

• Digital computation is too rigid because digital
  bits are either on or off
• Biological intelligence has subtle, analog,
  gradations
• Response we can use digital controlled analog
  methods, like the brain does, in machines
• Digital computation can simulate analog to any
  degree of (in)accuracy
• Challenge Mistakes in biological computing lead
  to serendipitous discoveries, a key to innovation

6
Criticism from the Complexity of Neural
Processing

• The information processes in the interneuronal
  connections (axons, dendrites, synapses) are far
  more complex than the simplistic models used in
  neural nets
• Response Newer models and simulations capture
  non-linearities and intricacies of biological
  originals
• Brain regions are not always as complex as the
  neurons in them
• Models now exist for several brain regions
• The genome has 30-100 M bytes of design info that
  gives rise to multiple redundancies to give final
  brainthis is a manageable complexity for models
• Caveat living experience trains the naive brain
  to define the working set of neural pathwaysthe
  intelligent computer will need experiences or
  be trained (and be trainable)

7
Criticism from Microtubules and Quantum Computing

• The microtubules in neurons are capable of
  quantum computing, and such quantum computing is
  a prerequisite for consciousness.
• To upload a personality, one would have to
  capture its precise quantum state.
• Response no evidence for either statement.
• Quantum computing does not require biological
  materials
• Personalities spring from an ever-changing
  quantum state (Im different than I was a moment
  ago)
• But even if you used my status from a previous
  moment, the copy would pass the Turing test.
• Caveat Microtubules are in constant flux of
  polymerization, de-polymerization and movement in
  three dimensions. They are crowded with loosely
  attached and mobile vesicles, and a web of other
  cytoplasmic fibers. Not a good substrate for a
  computing platform

8
The criticism from the Church-Turing thesis

• We can show that there are broad classes of
  problems that cannot be solved by any Turing
  machine
• As Turing machines can emulate computers, or
  solve problems a computer can, then computers
  cant solve all problems, whilst humans can
  therefore, machines will never emulate human
  intellegence.
• Response Humans are no better than machines at
  solving insolvable problems.
• Human educated guessing can approach solutions,
  but machines can do this too, and faster
• Logic, mathematics and computation have limits
• There exist equal numbers of soluble and
  insoluble problems

9
The Criticism from Failure Rates

• Computer systems are showing alarming rates of
  catastrophic failure as their complexity
  increases.
• Thomas Ray writes that we are pushing the
  limits (of) conventional approaches
• Response We already have many highly complex
  systems that have very low failure rates in
  system critical tasks
• Imperfection is inherent in complex processes,
  and that includes human intelligence

10
Criticism from Lock-in

• Complex expensive infrastructure in such services
  as energy supply and transportation stifle
  innovation and rapid change
• Response Rapid paradigm shifts follow the
  marketplacecheaper, faster, better help move to
  acceptance of new technology
• The internet is an example, nanotechnology will
  folllow
• Caveat At the further future end of Kurzweils
  log-plot projections, some near-miraculous
  engineering is requiredand on a compressed time
  schedule

11
Criticism from Ontology

• John Searle and the Chinese Room a man following
  a written program answers in Chinese questions
  presented in Chinese, he doesnt speak or
  understand Chinese, but follows a program to
  guide him to correct answers (as does a computer)
• Response The argument is a tautology, Kurzweil
  shreds Searle using quotes from Searle (p459)
• Our neurons dont know the languages we speak,
  it is the complexity (pattern) of the whole brain
  that recognizes patterns and provides correct
  answers

12
Criticism from the Rich-Poor Divide

• Only the rich will benefit from the new AI
  technology
• Response As technologies develop, price comes
  down, this will happen with AI computation
• Example everyone has cell phones now (once only
  the well-to-do)

13
Criticism from Government Regulation

• The government will slow or stop, at least
  impede, the exponential growth of AI technology
• Response Government regulation has impeded
  little the technologies involved in the AI
  revolution
• Progress will rush around rules like stream water
  rushes around stones in a stream bed

14
The Criticism from Theism

• According to Wm. Dembski, materialists such as
  Ray Kurzweilsee motions of matter as
  sufficient to account for human mentality while
  ignoring the spirit behind consciousness
• Response Dembski looks only at current
  computers, not ones billions of times more
  complex and capable
• The rich complexity of the human mind, even if
  defined, still leaves us to wonder at its
  remarkable qualities
• Computers are beginning to use chaotic processes
  to generate processes that are used in
  pattern-recognition just as the brain does

15
The criticism from Holism

• Michael Denton writes that organisms are holistic
  in that the entire being is involved in function
  and reproduction through biological processes
• Response While biological design has a profound
  set of principles, machines can use the same
  principles to create the emergent properties of
  patterns as did biology

16

• Finis