Contrasting a human brain to a computer

His Divine Shadow · Post by **His Divine Shadow** » 2003-02-25 11:06am

I think you have misunderstood what is said in the book, it's probably just a metaphorical expression and not meant to be taken literally.

Kenny_10_Bellys · Post by **Kenny_10_Bellys** » 2003-02-25 11:14am

The interaction is amazing, I agree, and I do not think things are stored as a hard drive. The initial premise was to design an artificially intelligent computer, not axions. Mimic the software and you get your intelligence, entirely seperately from the brain or system it's running on. We need the software design that the brain runs, not the brain. You dont make an AI computer by cutting a guys head in two and grafting him into it, that would be a guy linked to a computer.

His Divine Shadow · Post by **His Divine Shadow** » 2003-02-25 11:28am

Kenny_10_Bellys wrote:The interaction is amazing, I agree, and I do not think things are stored as a hard drive. The initial premise was to design an artificially intelligent computer, not axions. Mimic the software and you get your intelligence, entirely seperately from the brain or system it's running on

Software being the overall pattern of all the interactions between the neurons, one could potentially record this, but not with todays tech anyway.

We need the software design that the brain runs, not the brain. You dont make an AI computer by cutting a guys head in two and grafting him into it, that would be a guy linked to a computer.

If it's based on an adult human, that is the layout of the neural connections, it'll actually come equipped, in that sense, it'd be something of a duplicate.

However, if we make a baby-brain, it could be "raised", like a human would be raised, reaction to stimuli essentially, which makes neurons fire, some bonds strenghtening, some weakening, physical alteration of the layout of the neurons in other words.

His Divine Shadow · Post by **His Divine Shadow** » 2003-02-25 11:30am

Anyway, I feel the most interesting aspect of this, is the possibility of applying it to human brains, see the note on immortality or enhancements I made earlier.

Post by **NecronLord** » 2003-02-25 01:17pm

His Divine Shadow wrote:
NecronLord wrote:Not the best way, the most advanced way. The problem is maintainance, with option 2 maintainance becomes fairly easy, as every few minutes the AI is backed up to non-volitile storage, with option 1 a good kick screws it.
Exactly whats your proof of that? A human head can take a kicking and come out of it without problems.

With luck, just as often it's screwed in some way...

And the tech level of such a brain would require advanced micro-tech on the cellular scale, "repair/maintenance-cells" is not impossible.
And this "computer", or rather, it's neurons and axons would likely be suspended in some fluid of some form and be quite "mushy" or jello like, not hard like a CPU or circuit board.

I was talking about "option 2", you are talking about "option 1".

In any case I prefer the old computer axiom of KISS, Keep it simple stupid. If you can use option two, then it is inherently superior to option 1, as it is far easier to maintain, and can of course be upgraded easily. repair maintenance cells would a- take a long time to do anything b-require massive leaps in technology to be able to put things back the way they were even if they are possible, which is unlikely, the human brain cannot self repair any major damage, I find it unlikely that this could. Where would they get raw materials from? whould they have a miniture map of the AI to work from? etc etc.

His Divine Shadow · Post by **His Divine Shadow** » 2003-02-25 03:23pm

With luck, just as often it's screwed in some way...

Nope, it's not, it's extremely seldom that you suffer any brain damage from a few bangs to the head, infact you can survive pretty heavy shit and survive, it's pretty resilient, it's hard to smash things that small, especially when suspended in a shock absorbing gel like matter.

In any case I prefer the old computer axiom of KISS, Keep it simple stupid. If you can use option two, then it is inherently superior to option 1, as it is far easier to maintain, and can of course be upgraded easily.

It's unlikely it'll ever become mobile though.
I'm not talking about a sentient house, thats not interesting, someting mobile please.

repair maintenance cells would a- take a long time to do anything

Why? Nevermind, suffice to say thats an unprovable assumption.

b-require massive leaps in technology to be able to put things back the way they were even if they are possible, which is unlikely

Nope, if you got the tech to build those things, you got the tech to build maintenance devices too.

the human brain cannot self repair any major damage, I find it unlikely that this could.

This is not the human brain, I find it likely that it could, see, I can just make stuff up too.
And don't underestimate the human brain either, and especially not the inorganic equivalent, the problem with largescale cell death in major brain damage is due to lack of oxygen to cells, inorganic cells though, would run on electricity I assume, and would not decompose.

Where would they get raw materials from? whould they have a miniture map of the AI to work from? etc etc

What the fuck does this have to do with anything? Do I have to name every fucking last possible variable between the birth and end of the universe?

I mean, what do you want, should I just name things out of thin air and say, this will do that and this will do blah blah blah....

Why don't you start naming every component in and device that'll be involved in your scenario and the names and lives of the people who will make it?
Ask something that can be answered within any reasonable degree.

The general gist would be pretty frickin' obvious, it has a body, we're not making a self-sustained brain in a totally cut off universe, the brain can be divied into local sections too that are regularly checked and important info stored, but I'm just making shit up here on how it could possibly work, but it's what you where asking for.

Pu-239 · Post by **Pu-239** » 2003-02-25 03:56pm

Option two is better, with software running on a cluster, dividing the function of each node amoung the computers on a cluster. Building the nodes physically would be a pain in the ass, unless you can physically directly connect a node to hundreds of others, and rewire them on the fly. Nanobots would increase complexity exponentially. Putting in stuff for routing would increase the complexity also, since you would have to store addresses and not only states. Then again that might be within our reach. Software performance would be terrible, however if you are going to grant control of something to a self aware AI, the software could be programmed with failsafes, unless you are going to make the software design a better version of itself, like what they are doing now using evolutionary techniques to design simple circuits by modeling random circuits, then picking the ones with desired attributes, and further alter them. Of course designing something as complex as itself would take a shitload of time. I don't know, I know nothing about the brain or neural networks.

His Divine Shadow · Post by **His Divine Shadow** » 2003-02-25 05:04pm

I'm proposing that one uses a human brain as a template.

His Divine Shadow · Post by **His Divine Shadow** » 2003-02-25 06:05pm

Still, one could use a mix of option 1&2, small computer nodes that say handles the equivalent of a 10000 neurons on software, with that you could lower the amounts of physical neurons from 100 billion to 10,000,000 a much more manageable figure, this would make it less complicated by a large factor, whilst making it easier for software backups to another place, this allows for easier backups and maintenance whilst not running it all as software(which would create alot of overhead).

This might even be possible to cram into the rough dimensions of a human brain some time in the future.

Pu-239 · Post by **Pu-239** » 2003-02-25 07:40pm

His Divine Shadow wrote:Still, one could use a mix of option 1&2, small computer nodes that say handles the equivalent of a 10000 neurons on software, with that you could lower the amounts of physical neurons from 100 billion to 10,000,000 a much more manageable figure, this would make it less complicated by a large factor, whilst making it easier for software backups to another place, this allows for easier backups and maintenance whilst not running it all as software(which would create alot of overhead).

This might even be possible to cram into the rough dimensions of a human brain some time in the future.

That's why I said cluster of computers.

A lot of stuff is just random connections, which become organized as it learns and aquires memories.

UltraViolence83 · Post by **UltraViolence83** » 2003-02-25 08:53pm

His Divine Shadow wrote:Anyway, I feel the most interesting aspect of this, is the possibility of applying it to human brains, see the note on immortality or enhancements I made earlier.

False immortality is for losers and posers.

GO ELVES!