On Feb 23, 5:44*pm, "Rod Speed" wrote:
casey wrote
[...]
What machines can do better than us is: Carry out tasks
we give them at great speed which means they can do
things we can't because we can't write that fast and
we at this stage can't rewire our brain for those tasks.

They can also do tasks which require much more
reliable memory than any human can have too.

There may well be a reason we don't have such computer
perfect memories.


That's why navigation system do much better than any
human can ever do, they can use immense databases
which produce a much better result when say producing
an optimum route to cover a very wide variety of places
that must be visited with say a complex delivery run etc.

But all that was coded by humans and there may well be
a reason we are not that "perfect". Car navigation systems
have failed in a way a human would probably not fail
leading the drivers into dangerous situations.



Also our neurons are slow compared with electronic
switching devices which is a physical limitation not
a cognitive limitation. We can process visual data very
fast because we have parallel wiring for that task.

But it remains to be seen if we can do much better
with a machine that say uses measurement for facial
recognition when the image quality is poor etc.

Machines may well do better than humans but again it
depends on how clever we are at writing their code
unless of course new connections are learned in an ANN
of some kind. TD-Gammon is a good example of an ANN
that learnt to recognize high value backgammon board
states.


So let us not confuse speed of execution with intelligence.

Sure, but not all computing is about speed of execution.

And I wasn't suggesting otherwise. It is however the reason
we use computers for what they do so far is mostly what we
code them to do and which given their speed we could do also.
They don't write their own code or have their own goals.