Hey there, comp sci IQfyentooman here. I recently wanted to know what it actually is you math nerds do all day and I was quite shocked about what I learned.

Apparently mathematics is all about prepositional statements and truth values? And the goal of an mathematicians is to proved theorems as true or false?

I laughed so hard when i learneded this. You math geeks are a relic of old times, quite frankly you're antiquated and technology already replaced the need for you.

Don't belive me? I'll explain. There's this invention called PROGRAMMING and we use PROGRAMMING LANGUAGES. Hopefully I'm not moving too fast for you here. But basically, us comp sci GODS can easily decide what theorems is true oar not.

Observe:

Int main(void){

Boolean truth_Value = True;

pythagorean_Theorem = truth_Value;

Return 1;

}

Its that easy. No stupid proofs needed or any dumb stuff like that. Only a caveman would waste time crunching numbers and stuff when such a sleek, elegant, sophisticated solution already exists. Sorry, but you wasted your time getting an math education. If you want ill teach you how to coding so you can into having a REAL job.

DMT Has Friends For Me Shirt $21.68 |

DMT Has Friends For Me Shirt $21.68 |

>error Undefined symbol 'pythagorean_Theorem' line 3

It was predefined earlier with a #macro

Define it here for us

You're barking up a tree that Kurt Gödel went up almost 100 years ago. I hope you meet the same conclusion he did. Bear in mind, he didn't require a physical computer to do it.

>Apparently mathematics is all about prepositional statements and truth values?

yes, and it always has been. Finding the fundament behind patterns and rules, so that it can then be applied in practicalities like finance and engineering.

It's easy to talk when you have computers and all this amassed knowledge; you should try achieving progress from the very beginning, with zero-knowledge of anything, with sticks and stones, like our ancestors did.

Good luck.

>you should try achieving progress from the very beginning, with zero-knowledge of anything, with sticks and stones

Not possible. Can't unlearn what is already known

Most programmers will be replaced by AI and pajeets in the next decade but mathematicians will exist forever.

The future is now, old man.

Low quality low IQ bait

I disagree

Wrong. Watch Dr. Tao's recent talk on computational proofs. Your days are numbered

Construct a language, than you can call yourself a CS god, until then you are just a code monkey

Great, so I assume you can explain to us the fundamental combinators of Lambda Calculus, which all programming languages are based on:

[math] I = lambda x.x [/math]

[math] K = lambda xy.y [/math]

[math] S = lambda xyz.xz(yz) [/math]

[math] D = lambda x.lambda y.(x y)(y x) [/math]

[math] Y = lambda f.(lambda x.f (lambda v.x v) (lambda v.x v)) [/math]

[math] Z = lambda f.(lambda x.f (lambda v.x v) (lambda v.x v)) [/math]

[math] Omega = lambda x.xx [/math]

Oh, it appears I made a mistake here because Y and Z are not supposed to be the same. Can you please tell me how to fix it?

What is the purpose of D ()whih isn't even typable?

From my understanding,

-the other K' (actual K):= x,y -> x;

-B:= x,y,z -> x(yz)

-W:= x,y -> (xy)y

-C:= x,y,z -> (xz)y

Are more useful, in terms of Curry-Howard correspondence and its extension to classial logic (a.k.a. classical realizability)

LOG(X)+LOG(Y) = LOG(X*Y)

LOG(3) + LOG(4) = LOG(12)

LOG(5)*2 = LOG(5^2)

LOG(5)/2 = LOG(SQRT(5))

Ergo, Log-space is exponential space. Algebra-Space can do normal addition & subtraction.

Why is this info useful? Because neural nets use multiplication and division like crazy.

Convert your neural nets WEIGHTS database to logarithmic format and you reduce billions of calculations to thousands of cheap additions & subtractions. Granted you will have to do logarithmic conversions to allow basic Algebra-Space true additions & subtractions, but the costs of these conversions are dramatically CHEAPER than the billions of previously done multiply & divide operations. There are a few tricks to get true additions & subtractions in LOG-SPACE, but it beats any matrix math choices you can create.

>you will have to do logarithmic conversions to allow basic Algebra-Space true additions & subtractions, but the costs of these conversions are dramatically CHEAPER than the billions of previously done multiply & divide operations.

Maybe, maybe not. Maybe the chip designers are already doing it, maybe not.

You could write an RTL implementation of this concept and show how it compares to "regular" floating point multiplication and division, whatever the state of the art is. Also, wrong thread.

As far as I'm aware the reason people use log-sum-exp for weight updates/renormalization is to avoid numerical precision problems. I don't think it really has much of anything to do with saving computation time and is more to avoid underflow for low probability weighted events in the weight update procedure.

yes math was outdated with my birth ... now the adventure leads the show -

i could see your logic i suppose ... but i disagree with your lack of a barrier around each cell you didn't really excel to me ...

why can some people not exit buildings?

>code monkeys write shit this and think they're profound