Moravich second half:
his timeline is certainly too optimistic
the bush seems like a good idea, though Moravich never published it in
a scientific journal
What are his assumptions?
- computers will get cheaper and more powerful
- so we can assume we will be able to do anything we can
- machines will become more powerful than people--he
firmly believes this
- the human mind is mechanistic
- there is no soul
moving humans into robots in the least palatable
if the human mind is mechanistic is there no free will?
he seems to assume we will move towards combining our minds--aren't we
more individualistic than that?
but science has become more group research (distributed cognition)
is the end goal to create something as close to human as possible or
something better than humans
can robots innovate?
How would you make the argument on the other side--that
computers can never do what humans do
- maybe not by genius
- but by trial and error, which is very effective
- would computers all come up with the same result?
- you can use an evolutionary process
- Moravec wants to start with lower functions and keep
- is human inspiration different from what you can do on
- is everything we do based on physical processes in the
either we try to make computers like humans or do we make
them more efficient than humans
- we aren't substantially closer to an innovative
computer--but experts would probably disagree
- is there a fundamental jump that would need to be made?
- robots don't have touch very effectively--they don't
have the sensory experience which maybe is the basis of brain
- is matching human intelligence really what we want--why
slow down the computer to match a human
- can a computer have judgement?
- we have different experiences and values and intrinsic
qualities--those can't be emulated
- if you reproduced the brain it still wouldn't be the
same as human thought
- researchers assume: everything we do come from
physical processes in the brain
- if we reproduce a complete set of physical processes
then magic it will be able to think
- counterargument: different materials have different
needs and so the different basis will inevitably have different results
- can you program a computer to rewrite itself? how
completely?--people are trying
- you can't fit our environment without being like us,
but our environment only works for us, it will never work as well for
something else we create
if they can do the job better than people then do we reduce the role
is human action really just programmed reactions
if it works the same it is the same?
could we mimic animals with a robot?
how about gathering the information in someone's brain and sharing it?
the internet gives us some of that
what about free will?
- Moravec thinks robots can have free will too
- do we want computers/robots that can make mistakes?
- overriding rules for robots take away free will but can
What is AI?
- pass the Turing test
- Deep Blue