How would you solve this

[chegra]

Lets says you have a 1MB of data in binary:
000010010001010......000110001110

Now, Lets say you have all the premutation of 1MB:
1. 000010010001010......000110001110
2. 111000101000010......000110010100
3. 010010100101010......000000000000
:
:
2^8,000,000. 00011000100......1000001010

Now if you noticed, that if you had all the permutation of 1mb
basically, if the binary data was used to represent pictures it would contain
all the possible pics that could every exist under 1mb(yes your universe is finite, atleast to a certain degree).

If it was used to respresent executable it would contains all the possible programs that could exist under 1mb and if you apply the same logics for mp3 all the possible musics under 1mb.

Now, if you think about it, it will contain all the cures for any disease that can be stored in a word document that is under 1mb.

Well we could look at each permutation one by one to find what we are looking for, but it will take several life times of the universe to find exactly what we are looking for.

So, basically I am asking how would you go about solving this problem not in linear time but in constant time.

 

[Fluffywolf]

Well, a cure to a disease is more then something you can put in a file. Suppose, based on the same assumptions you create a textfile with every possible explanation for a cure for cancer.

Suppose the file says

The Cure for Cancer is <every possible answer>

How would you know which of the possibilities is the cure. And how would you know how to apply that solution.

And this is just for making all possible text files. If you create all possible programs based on machine code. Just seems like a waste of time. The concept is pretty cool though. Still, 99.99999999999<and possibly a lot more 9's>% would do nothing but error out. And the smallest of percentage that does will still have so many options, the chance for something valuable that hasn't been 'invented' the conventional way is astronomically slim. You might get lucky and run into some decent porn though.

 

[chegra]

If a human was doing by hand lets say a chemist, who has infinite life he would be able to test each one.

So it could be solved brute forced, but what I am looking for is an inefficient solution

 

[chegra]

I Particular post this in NT commune cause, its more of NT thing.

And generally, I wouldn't have to explain why it would work and basically focus more on generating ideas for solutions.

I think NT would be intrigue by this sort of thing.

 

[chegra]

For NT, if you dont recognize, it is the british museum algorithm

 

[Fluffywolf]

Well, I see another problem, that's that, because you don't know the cure for cancer(I'm sticking with this example but you understand it'd be the same for any example). You have no idea how to write a diagnostic tool to sniff out potential cures.

So even through running them in a protected shell, with a diagnostic tool sniffing all over. There would be know way of telling the diagnostic tool what to look for.

I mean, there will be 70368744177664 programs that utilize 1 Mb as effective code.
Take one bit out of the picture. So 1Mb - 1 bit and you have another 70368727400449 programs to scan through. And so on and so forth.

So apart from brute force which really won't work. You'd need a highly sophisticated diagnostic tool. A tool that somehow can sniff out the answer of the problem you seek. But in order to make a diagnostic tool to be able to sniff out the answer, you already need to know all of the knowledge that went into the question.


Now, if we ever make a simulation that can do chemistry as if it was real time based on the very nano building blocks matter is created with. And let a simulation like that brute force simulate through all kinds of formulae. All the while have a simulated piece of cancercombined with healthy flesh, brute force simulate with a super computer and study the positive results that removed the cancer leaving the healthy cells in tact.



So, create a simulation capable of calculation the right question, and your idea might have great merit.

 

[chegra]

Now, if we ever make a simulation that can do chemistry as if it was real time based on the very nano building blocks matter is created with. And let a simulation like that brute force simulate through all kinds of formulae. All the while have a simulated piece of cancercombined with healthy flesh, brute force simulate with a super computer and study the positive results that removed the cancer leaving the healthy cells in tact.

Ok, I didn't want to put my solution cause I didn't want it to influence the thoughts of others.

But you are right about creating a tool to test.

Firstly, to get the the calculation done in constant time it needs to be done in parallel.
So we will need a machine that could perform 2^8,000,000 test all in one go.

The only thing, I think that could do it is a quantum computer which can perform 2^L calculations with L qubits in constant time, but the current state of the technology is only 5 qubit, so it might be a while before we get to 8,000,000 qubits.

Next I would run a algorithm of a strong AI on the quantum computer, something that has yet to be achieve but conceivable in the next 50 years.
The strong AI would basically have the brains of a chemist.


My next solution would be to put a human in a superposition state and manipulate his wavefunction to collapse in a state where he knows the answer. Problem - Hard to put a macro-organism into superposition and manipulate.

But, I was looking for solutions that doesnt incorporate quantum physics.

I think the first one could solve it, maybe in the next 100 years or so.

 

[chegra]

But instead of cancer, lets say music.

Since it doesnt take extra tools to identify good music.

 

[Orangey]

Moved from The NT Rationale.

 

[Blank]

I would have someone else do it. =D

 

[EcK]

the best pattern recognition software\hardware mix ever (alot of them) to read the outputs and quantum ("classical or gravity) processing for the core unit.

It's an old idea btw, scifi authors such as stanislas lew wrote about it. And we all thought about it at some point.

But it'd take more time to find paterns than simply do the research.
It's like setting a computer to go random and wait until it wrote the best book ever.
Even without starting about the subjectivity of the thing and the sort of patern recognition ai we'd need to dev. it'd be easier to simply write the damn thing.

 

[matmos]

the best pattern recognition software\hardware mix ever (alot of them) to read the outputs and quantum ("classical or gravity) processing for the core unit.

It's an old idea btw, scifi authors such as stanislas lew wrote about it. And we all thought about it at some point.

But it'd take more time to find paterns than simply do the research.
It's like setting a computer to go random and wait until it wrote the best book ever.
Even without starting about the subjectivity of the thing and the sort of patern recognition ai we'd need to dev. it'd be easier to simply write the damn thing.

And it's useful if you wish to go data mining.

 

[chegra]

the best pattern recognition software\hardware mix ever (alot of them) to read the outputs and quantum ("classical or gravity) processing for the core unit.

It's an old idea btw, scifi authors such as stanislas lew wrote about it. And we all thought about it at some point.

But it'd take more time to find paterns than simply do the research.
It's like setting a computer to go random and wait until it wrote the best book ever.
Even without starting about the subjectivity of the thing and the sort of patern recognition ai we'd need to dev. it'd be easier to simply write the damn thing.

At first it is easy to simply write the damn thing, but I think if we can solve it in constant time, it will present solutions that are non-linear.

 

[chegra]

And if you think about it, humans, are just an efficient searching algorithm for this search space.

 

[EcK]

seriously, where are you getting at.
The problem in itself is stupid.
If you could have the technology to make it doable in human scale time and don't start to play around with relativity too much (like leaving in a space ship and coming back a zillion years later like 'hey, it solved the problem in 10 seconds of ship time! (and please don't start about acceleration\deceleration)

So the point is, that if you had that technology, you'd also be able to pretty much compute anything, inputing some basic information, adding not only patern recognition ias for the research, but also some goals and a set of rules+ a patern creation thing to create new data and connecting it to the existing data if needed.
It'd be more efficient by a ridiculous amount of orders of magnitude.

 

[EcK]

and you're trying so hard to look smart that it isn't even funny.
Saying obvious things using excessively technical terms for no good reason is something idiots do.

'm just saying

 

[chegra]

seriously, where are you getting at.
The problem in itself is stupid.
If you could have the technology to make it doable in human scale time and don't start to play around with relativity too much (like leaving in a space ship and coming back a zillion years later like 'hey, it solved the problem in 10 seconds of ship time! (and please don't start about acceleration\deceleration)

So the point is, that if you had that technology, you'd also be able to pretty much compute anything, inputing some basic information, adding not only patern recognition ias for the research, but also some goals and a set of rules+ a patern creation thing to create new data and connecting it to the existing data if needed.
It'd be more efficient by a ridiculous amount of orders of magnitude.

Let me see you are saying the problem is stupid because if we could that means we could do it a better way?

lol

The point of the question is to elicit a solution that doesnt require advance solution.

 

[EcK]

See previous post

 

[EcK]

There is no such solution. You can't magicaly 'find the solution you give a meaning to' out of simple 'neutral' data. Maybe if you accept the idea of a god, but it's theoreticaly unsupported and a bit simplistic.

Ultimately we just find logic to be 'logical' because it fits our internal neural paterns. And that logic is relevant and applicable to this universe because our brain gets constant feedbacks from said universe and has evolved as an body(brain included)\universe interface.

You'd need to at least partially simulate a or a set of human brain(s) working at much higher processing speed to find the paterns ur interest by(so that was what the ai patern recognition things were about). But as I said previously, it's as energy and time efficient as blowing up the universe to open a can.

Let me see you are saying the problem is stupid because if we could that means we could do it a better way?

yes

 

[chegra]

and you're trying so hard to look smart that it isn't even funny.
Saying obvious things using excessively technical terms for no good reason is something idiots do.

'm just saying

lol,
Why the personal attacks?

Generally, I dont like posting stuff about what i think because
some how it gets into some kinda personal thing.

Is it relevant if I am a idiot or not to this discussion?

If you think the problem don't have any merit then simply don't comment.