Agreed that you never claimed “mysterious or magical”. But I’m a little confused on what you mean by “algorithmic” in this case. If you accept that “algorithm” is equivalent to “computable”, then the statement “the brain does not operate algorithmically” is equivalent to saying that A.I. (at least human-type intelligence implemented on computers*) is impossible. For clarity’s sake – is that what you mean?
– I left this as “computers”, although perhaps it should be “digital computers” or “Von Neumann architectures”. I believe that, in this case, the term “computers” is sufficiently expressive.
Then what you are saying is that the proclamed limit is the time we have on earth, because if we had unlimited time we might have be able memorize the Internet (say in a million years), I’m sure. Also, your point concerning the conscious/unconscious is of course speculation (though having read about twenty volumes or more of Freud’s and Jung’s collected works - no joke - I’m inclined to agree). I was only wondering from what standpoint you made your statement that learning and self awareness [aside of time] is limited. I’m still unsure myself, but the again, I usually am.
First, and most obviously, we only have so many neurons. Second, our memory is imperfect; we forget things, or distort them. Third, there is a great deal of research showing that a great deal, probably most of our mental function don’t come to our conscious attention, and that what we are “aware” of is often illusion ( I wouldn’t consider Freud or Jung relevant ).
A classic example for the illusion part comes from split brain patients ( whose left and right hemispheres are severed from each other ). The right hemisphere can understand simple words and commands on it’s own; whisper “walk” in the ear it controls and the person will get up and walk. Ask him why did did that, and he’ll say something like “I wanted to get a drink”; some unconscious part of his mind made up an explanation for him, and it feels real to him.
As for memorizing the internet or anything else over a million years, I expect you’d just forget things as fast as you learned them after a certain point. The brain is designed to forget things, apparently.
That wasn’t the jist of the discussion though. It started with Lemur866’s post:
I was trying to nudge the discussion in the direction of whether it is true that a brain is qualitatively different from a computer in such a way that not only would it be difficult to achieve AI using computers as we know them, but that it would in fact be impossible. I suspect that it is not impossible, and is a matter of degree rather than kind.
Sorry I just don’t buy this - this is the Urban myth of Bit Rot and the like. Saving on a HD does not work that way, unless something is changing the file (i.e. virus or some such). Files saved on the HD are static, not dynamic, and once the compression algorythm is executed it does no further changes to the file. For example - take a digital photo, save it as a JPG. It will do the 10-1 compression by chopping out parts of the photo, but it only does it once. Next time you open the photo it will not change.
Computers don’t work that way. Storage media can degrade (i.e. magnetic tapes can loose their magnetism and data can get lost) but the data itself does not change once it’s stored to a HD unless something changes it.
Hear, hear! Just as a fun exercise for anyone skeptical of this point, try the following:[ol][li]Pick an arbitrary file. (it can be a jpeg if you want)[]Generate an MD5 checksum for that file and squirrel it away somewhere safe.[]Open the file.[]Close the file.**[]Generate a new checksum for the file and compare it to the first one.[]If the two checksums are identical, repeat steps 3 through 5 until they aren’t. If they’re different, let us know how many iterations it took.[/ol]…get back to us when you’re done. [/li]
For additional fun:[ul][li]Insert another step (“5.5” perhaps) in which you turn off the computer and ignore it for some arbitrary amount of time in the middle of each iteration.[]Without actually modifying any particular part of the file’s data by hand, click the “save” button as many times as you like before closing the file. (If this ever actually does anything, you’ve a sketchy editor or a file format that incorporates a timestamp in the file header.)[/ul][/li]**Note that we’re just opening and closing the file. We not editing/saving the file again or subjecting it to additional compression.
Mostly correct, but I know of cases of silent data corruption, where, due to a bug in the CPU, data being written is not what is expected. But only hardware failures should affect data on the disk - unless the read or write heads are bad, or there is a fault in the electronics behind them, or in the processor on the disk or …
(I spend my days analyzing failures, so I’m a bit paranoid…)
Like I said - the file only changes if something changes it. It won’t change itself, just by opening and closing, or over time, unless some other process changes it.
But since you can’t observe the state of the data without going through the read electronics, to a first approximation a fail on the disk caused by dust, or degradation of the magnetic surface, is indiestinguishable from other fails. It takes some debug to find the root cause.
It is true that things fail for a reason - but to the plain user, reliability fails look a lot like bit rot.
[/li]