/sci/ - Science & Math

[Anonymous No.15909770]

Math is just memorizing a bunch of arbitrary rules and formulas and the people who excel the most are the ones with the best memory. What a stupid fucking field of study. Why even learn it when you will never be able to outperform a computer at it?

[Anonymous No.15909773]

>>15909770
Computers got to be programmed.

[Anonymous No.15909783]

They are STILL trying to find a computer that can solve mathematical olympiad problems as well as a human can.

Good luck explaining to a computer what a nonstandard model of ZFC is.

Good luck teaching a computer how to invent novel mathematical structures essential to proving desired results.

[Anonymous No.15909813]

>>15909770
Cool, guess you're right.

[Anonymous No.15909892]

>>15909770
What isn't? Even sports is just best muscle memory and capacity.

[Anonymous No.15910041]

[Anonymous No.15910094]

>>15909770
that's why you use a computer to go fast, and if you don't you're peak retard

[Anonymous No.15910121]

>>15909770
Based on your post I can only conclude that your knowledge of mathematics is at best that of a middle-schooler, since you seem to think it's just arithmetic and elementary algebra.
How do you think the technology used to build computers was created? How do you think the algorithms that make them so fast were invented? Did computers do that on their own without human help?

[Anonymous No.15910126]

>>15910121
Computers have nothing to do with math beyond basic arithmetic and boolean logic.

[Anonymous No.15910135]

>>15910126
Not him but what math can't be decomposed into that?

[Anonymous No.15910148]

>>15910135
>>15910041
I've heard more than once it would take 100 mathematicians 10 years to program the proof of Fermat's last theorem into a computer to double check it

[Anonymous No.15910168]

>>15910126
Algorithms, the chemestry and quantum mechanics used to build better hardware, the tools needed to program them to be able to solve engineering problems, etc would disagree.

[Anonymous No.15910179]

>>15910148
I'm asking what math can't, not what math can.

[Anonymous No.15910186]

>>15910126
>>15910135
Math doesn't have a solid foundation in first-order logic.

https://www.youtube.com/watch?v=mhIkyqLDl9M

[Anonymous No.15910225]

>>15910186
Irrelevant to the question of what doable math can't be done on a computer.

[Anonymous No.15910271]

>>15909770
>he memorizes things instead of deeply understanding it
ngmi

[Anonymous No.15910395]

>>15910135
For starters anything with (large) sets, since a theory that only has arithmetic and logic is way smaller than ZFC. A more pragmatic example is anything that uses Axiom of Choice (or any of its variants) at some point, and Fermat's Last Theorem definitely does.
Even if everything would come down to arithmetic and logic it doesn't make it easier. As a part of Gödel's incompleteness theorem there is an algorithm for checking validity of certain logical statements, when expanded it is clear that it takes an astronomical amount of operations before halting, and thus it's no useful. At some point, by using linear forms in logarithms, Catalan's conjecture was also true asymptotically, but the bound to check was just too large for computers to handle.

[Anonymous No.15910731]

>>15910135
>>15910395
Yes, pretty much all math statements can ultimately be broken into a series of boolean statements. What’s the problem now, shouldn’t the computer have a field day, then? No. Computers are still useless for actual non-arithmetic math problems. This is because the complexer and thus more interesting a statement, or its possible answer becomes, the longer it becomes converted into the form of a boolean expression (these are called "normal forms"). But with additional atom, you add one more possibility too the chain within the space of its permutations. The size thus quickly grows in the order of 2^n, nearly the worst time complexity. It’s completely, and with that I mean computationally/mathematically, not merely technologically, infeasible to try to bruteforce interesting math/other such endeavours.
You need intelligence to cut down this stack.
>so can AI do it?
Probably, machine learning simulates intelligence after all. But that is completely unrelated to the issue of computers not per se being able to solve non-arithmetic math. You could envision a machine learning system that you have bootstrapped from a couple trained circus dogs jumping through hoops in the physical world. It’s a bunch of regressors and convolutions, not inherently tied to computers.

[Anonymous No.15910752]

>>15909770
That's funny, that's not what math is about at all
the substrate of mathematics is merely the paint which is used to make art.

[Anonymous No.15910905]

>>15909770
Math *can* be, and usually is, that. However, you can still learn mathematics as theory and famous problems rather than as rote memorisation and calculation.

[Anonymous No.15910912]

>>15910905
You can learn by simply wondering about circles and triangles you god damn retard

[El Arcón No.15910924]

>>15909770
>never be able to outperform

[Anonymous No.15910968]

>>15909770
not all functions are computable, and not every computable function can be realistically executed on a finite tape
>>15910752
based