Re: [crit] Fallen

On Fri, 29 May 2009 11:58:21 +0100, Gerry Quinn <gerryq@xxxxxxxxx>
wrote:

In article <tqun1517c0crs8lt225ljkf487fk42pp3s@xxxxxxx>,
email_addr@xxxxxxxxxxx says...

Glad you've got that covered, though I do find it odd that someone who
works in AI would consider typed data to be a good thing... until the
issue of typeless data is Truly Handled, I think real AI will remain
purely theoretical and we'll be stuck with a lot of fancy
implementations of not-AI, most of which are not-real-smart.

Whether data is typed or not is a programming issue and nothing to do
with how the AI operates, which is on a different level. You might as
well argue that human intelligence is crippled by the fact that the
brain uses proteins and sugars as a basis for all operations.

Or that computers can't be intelligent because they *don't* use organic
chemicals. Alas, some people do make that argument. -)

There's also a similar argument at a higher level (but still way below
your "how the AI operates" level, which I agre with): there is a
discrete v. continuous distinction,. I don't *think* I've come across
anyone making this argument seriously[*], but if they did, it would be
something along the lines that digital computers can only be in a finite
number of states, whereas the human brain involves lots of continuous
processes, and therefore is capable of an infinite number of states.

Which reminds me: as a schoolboy, I had a six inch slide-rule. (I still
have it.) I had a lot of difficulty get 3 digits of precision out of it.
I rather envied my classmates with their 12 inch slide-rules.

Now, I have a cheap, scientific calculator which displays 10 digits, and
seems to work to 12 digits of precision. But it's digital.

Jonathan
[*] Possibly because "nerve impulses" are discrete. But that's a
red-herring, because it's the *rate* of firing which matters, and that
is a continuous variable. And the signalling isn't where the computation
takes place anyway.

--
"If I let everyone I thought was an idiot die - there
wouldn't be many people left." - Gilgamesh Wulfenbach
.

Relevant Pages

  • Re: [crit] Fallen
    ... works in AI would consider typed data to be a good thing... ... Whether data is typed or not is a programming issue and nothing to do ... Or that computers can't be intelligent because they *don't* use organic ... I had a lot of difficulty get 3 digits of precision out of it. ...
    (rec.arts.sf.composition)
  • Re: huge number calculator and library
    ... and 72 bit double precision. ... we also had a research lab that used analog computers. ... >>digits to 31 digits a question arose as to whether or not anyone KNEW of any ... why should the compiler raise an objection? ...
    (comp.lang.cobol)
  • Re: Calculating Wishes (was hpcatalog.com)
    ... But compared to modern computers, ... Sun-Earth distance useful even if it's know to a precision clearly ... digits of precision. ... > If a calculator offerred say 19 or more digits of precision, ...
    (comp.sys.hp48)
  • Re: This calculation is just wrong / computer cant count!
    ... calculators and computers are using the same kind of arithmetic? ... essentially carry integer numerators and denominators symbolically through ... precision. ... until you ran out of digits. ...
    (microsoft.public.vc.mfc)