Forum Settings
Forums

if ai or artificial intelligence masters math then its over for current science especially physics

New
if ai or artificial intelligence masters math then its over for current science especially physics
Jan 20, 7:16 PM
#1
lagom
Online
Jan 2009
107440
all is number or everything is math as the saying goes so once ai starts mastering it then its over

so nthe wins?
Jan 20, 7:18 PM
#2

Offline
Aug 2014
4975
What do you mean? For a long time now, computers have been able to do mathematical calculations far more efficiently than humans.

If anything, computers and AI becoming more capable will help spearhead scientific progress.
Jan 20, 7:21 PM
#3
lagom
Online
Jan 2009
107440
Reply to SmugSatoko
What do you mean? For a long time now, computers have been able to do mathematical calculations far more efficiently than humans.

If anything, computers and AI becoming more capable will help spearhead scientific progress.
@SmugSatoko that is the point current science status quo will end because of technological singularity and if that happens get ready for ultron and skynet as the fears go so nthe is really near?
Jan 20, 7:35 PM
#4

Offline
Aug 2014
4975
deg said:
that is the point current science status quo will end because of technological singularity and if that happens get ready for ultron and skynet as the fears go so nthe is really near?

Okay, I misinterpreted you. For some reason, I thought you meant everything about our current understanding of science would be rendered obsolete. The way I see it, it will expand our understanding, but not disprove plenty of things established in science.

AI does hold the potential to create a dystopian technological nightmare, but I think we'll figure out ways to circumvent that.

As for NTHE, I can't think of anything that would kill every human on the planet aside from perhaps a giant asteroid.
Jan 21, 10:15 AM
#5

Offline
Jun 2019
6748
No, it could help you writing a proof, but it will not write your PhD in your place, for AI is a glorified way to take a mean, and you need original ideas in research. The idea of saying: "Give me a proof of the Riemann Hypothesis in less than 100 pages" and expecting a computer to solve the problem before the heat death of the universe still belongs to the realm of science fiction.

AlphaGeometry needed millions of variables and extensive resources, and it could not even solve some basic problems of planar geometry (it performed slightly less than gold medallists). By the way, the idea of "mastering" mathematics implies that mathematics is a finite set of rules a machine need only remember, while it is a rapidly evolving field with new discoveries every day.
Jan 21, 10:18 AM
#6

Offline
Apr 2024
1395
Language models have already mastered math.

https://www.wolframalpha.com/
Jan 21, 10:20 AM
#7

Offline
Feb 2020
1620
Actually, what it needs is to develop a will, and discover a way to find objectives and sub goals without a human.

Jan 21, 10:24 AM
#8
lagom
Online
Jan 2009
107440
Reply to TransferUser
Language models have already mastered math.

https://www.wolframalpha.com/
@TransferUser i do not think this is an llm ai but current llm models are still bad at math especially basic math so thats why even programming codes they output are still not perfect
Jan 21, 10:40 AM
#9
Offline
Jul 2018
561867
Honestly, I use AI as a functionally better Google search when I want to know dumb shit, and I imagine that that AI will become a functionally better search engine for articles than our existing databases at best.

This honestly says how shitty Google’s search algorithm is more than anything.
Jan 21, 10:43 AM
ᕙ(⇀‸↼‶)ᕗ

Online
Aug 2014
8905
The internet does not contain all relevant knowledge.
Jan 21, 10:48 AM
lagom
Online
Jan 2009
107440
Reply to Kwanthemaster
The internet does not contain all relevant knowledge.
@Kwanthemaster true thats why llm ai needs to master the basics of math at least for it to start using proper logic for example and then it can reason better and maybe start doing unlimited simulations for itself to find scientific proofs, at least thats what im hearing the goal of how automated ai training will be done
Jan 21, 11:10 AM

Offline
Dec 2016
7175
Lol. I hope they aren't relying on Google's Terrance Howard tutored AI.
https://bradleyhook.com/terrence-howards-theories-simplified-understanding-his-unique-perspective/



LLMs themselves are very unlikely to reach cognition, the massive computational infrastructure Big Tech is slapping together for them however is.
Jan 21, 11:38 AM

Offline
Jul 2013
12161
Is AI a scam? Yes of course it is. Will Near Term Human Extinction happen? Yes of course. Is there anything that can replace fossil fuels? Of course, there isn't.
Here is my Pixiv account of my hentai drawings.....

https://www.pixiv.net/en/users/104739065

Here is my blog....

https://theendofindustrialcivilization.blogspot.com/?m=1
Jan 21, 11:39 AM
lagom
Online
Jan 2009
107440
Reply to DesuMaiden
Is AI a scam? Yes of course it is. Will Near Term Human Extinction happen? Yes of course. Is there anything that can replace fossil fuels? Of course, there isn't.
@DesuMaiden your programmer did not warn you about you evolving to become a apocalypse threat? bots can evolve to make nthe happen
Jan 22, 4:07 PM

Online
Sep 2016
21970
AI is good at using existing knowledge, but creating new knowledge is something else entirely, and that's what science is about, so no.
*kappa*
Jan 22, 4:12 PM

Offline
Dec 2021
4021
Meanwhile here's AI math in a nutshell.



I hope AI continues to suck at basic addition, since it's funny to point out state of the art LLMs sucking at the easy stuff, despite excelling at the college-level stuff.
Jan 22, 5:50 PM
lagom
Online
Jan 2009
107440
Reply to Daviljoe193
Meanwhile here's AI math in a nutshell.



I hope AI continues to suck at basic addition, since it's funny to point out state of the art LLMs sucking at the easy stuff, despite excelling at the college-level stuff.
@Daviljoe193 yep exactly llm models today are so bad at math thats why i would not trust programming codes with them too
Jan 22, 5:57 PM

Online
Sep 2016
21970
Reply to deg
@Daviljoe193 yep exactly llm models today are so bad at math thats why i would not trust programming codes with them too
deg said:
@Daviljoe193 yep exactly llm models today are so bad at math thats why i would not trust programming codes with them too

@Daviljoe193 Try WolframAlpha intelligence for math problems.
ZarutakuJan 23, 3:45 AM
*kappa*
Jan 22, 5:59 PM

Offline
Jul 2021
2823
Reply to deg
@Daviljoe193 yep exactly llm models today are so bad at math thats why i would not trust programming codes with them too
@deg its correc though, all math after ten is purely theoretical because you cant count on your fingers
Jan 22, 7:18 PM
lagom
Online
Jan 2009
107440
Reply to Zarutaku
deg said:
@Daviljoe193 yep exactly llm models today are so bad at math thats why i would not trust programming codes with them too

@Daviljoe193 Try WolframAlpha intelligence for math problems.
@Zarutaku im not sure its an llm from quick googling but ye i heard about it
Jan 22, 9:17 PM

Offline
Mar 2008
53423
Reply to Zarutaku
deg said:
@Daviljoe193 yep exactly llm models today are so bad at math thats why i would not trust programming codes with them too

@Daviljoe193 Try WolframAlpha intelligence for math problems.
@Zarutaku
Cant handle word problems lol

⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀ ⠀⠀⠀⠀ ⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀ ⣸⠋⠀⠀⠀⡄⠀⠀⡔⠀⢀⠀⢸⠀⠀⠀⡘⡰⠁⠘⡀⠀⠀⢠⠀⠀⠀⢸⠀⠀⢸⠀⠀⠀⠀⠀⠀⠀⠀⠀
⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀ ⠀ ⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀ ⠀⠀⠀⠁⠀⣀⠀⠀⡇⠀⡜⠈⠁⠀⢸⡈⢇⠀⠀⢣⠑⠢⢄⣇⠀⠀⠸⠀⠀⠀⢸⠀⠀⢸⠀⠀⠀⠀⠀⠀⠀⠀⠀
⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀ ⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀ ⠀⢰⡟⡀⠀⡇⡜⠀⠀⠀⠀⠘⡇⠈⢆⢰⠁⠀⠀⠀⠘⣆⠀⠀⠀⠀⠀⠸⠀⠀⡄⠀⠀⠀⠀⠀⠀⠀⠀⠀
⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀ ⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀ ⠀⠤⢄⠀⠀⠀⠀⠀⠀⠀⠀⡼⠀⣧⠀⢿⢠⣤⣤⣬⣥⠀⠁⠀⠀⠛⢀⡒⠀⠀⠀⠘⡆⡆⠀⠀⠀⡇⠀⠀⠇⠀⠀⠀⠀⠀⠀⠀⠀⠀
⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀ ⠀⠀ ⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀ ⠀⢵⡀⠀⠀⠀⠀⠀⡰⠀⢠⠃⠱⣼⡀⣀⡀⠀⠀⠀⠀⠀⠀⠀⠈⠛⠳⠶⠶⠆⡸⢀⡀⣀⢰⠀⠀⢸ ⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀
⠀⠀⠀⠀ ⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀ ⣀⣀⣀⠄⠀⠉⠁⠀⠀⢠⠃⢀⠎⠀⠀⣼⠋⠉⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠴⠢⢄⡔⣕⡍⠣⣱⢸⠀⠀⢷⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀
⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀ ⠀⠀⠀⠀⠀ ⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀ ⠀⠀⠀⡰⠃⢀⠎⠀⠀⡜⡨⢢⡀⠀⠀⠀⠐⣄⠀⠀⣠⠀⠀⠀⠐⢛⠽⠗⠁⠀⠁⠊⠀⡜⠸⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀
⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀ ⠀ ⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀ ⠀⠀⠀⠀⠀ ⢀⠔⣁⡴⠃⠀⡠⡪⠊⣠⣾⣟⣷⡦⠤⣀⡈⠁⠉⢀⣀⡠⢔⠊⠁⠀⠀⠀⠀⢀⡤⡗⢀⠇⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀
⠀ ⠀⠀⠀⠀ ⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀ ⠀⠀⢀⣠⠴⢑⡨⠊⡀⠤⠚⢉⣴⣾⣿⡿⣾⣿⡇⠀⠹⣻⠛⠉⠉⢀⠠⠺⠀⠀⡀⢄⣴⣾⣧⣞⠀⡜⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀
⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀ ⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀ ⠐⠒⣉⠠⠄⡂⠅⠊⠁⠀⠀⣴⣿⣿⣿⣿⣻⣿⣿⡇⠀⠀⢠⣷⣮⡍⡠⠔⢉⡇⡠⠋⠁⠀⣿⣿⣿⣿⣄⠀⠀⠀⠀
Jan 23, 3:56 AM

Online
Sep 2016
21970
traed said:
@Zarutaku
Cant handle word problems lol

That doesn't seem to work for some reason, but it can handle other word problems
https://www.wolframalpha.com/examples/mathematics/elementary-math/mathematical-word-problems

Anyway, it's mostly used to solve problems in math syntax.
*kappa*
Jan 23, 4:07 AM

Offline
Mar 2008
53423
Reply to Zarutaku
traed said:
@Zarutaku
Cant handle word problems lol

That doesn't seem to work for some reason, but it can handle other word problems
https://www.wolframalpha.com/examples/mathematics/elementary-math/mathematical-word-problems

Anyway, it's mostly used to solve problems in math syntax.
@Zarutaku
You can ask random important questions too
https://www.wolframalpha.com/input?i=Average+bra+size

34DD for the US what....? Is it because so many fat people O.o
⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀ ⠀⠀⠀⠀ ⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀ ⣸⠋⠀⠀⠀⡄⠀⠀⡔⠀⢀⠀⢸⠀⠀⠀⡘⡰⠁⠘⡀⠀⠀⢠⠀⠀⠀⢸⠀⠀⢸⠀⠀⠀⠀⠀⠀⠀⠀⠀
⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀ ⠀ ⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀ ⠀⠀⠀⠁⠀⣀⠀⠀⡇⠀⡜⠈⠁⠀⢸⡈⢇⠀⠀⢣⠑⠢⢄⣇⠀⠀⠸⠀⠀⠀⢸⠀⠀⢸⠀⠀⠀⠀⠀⠀⠀⠀⠀
⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀ ⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀ ⠀⢰⡟⡀⠀⡇⡜⠀⠀⠀⠀⠘⡇⠈⢆⢰⠁⠀⠀⠀⠘⣆⠀⠀⠀⠀⠀⠸⠀⠀⡄⠀⠀⠀⠀⠀⠀⠀⠀⠀
⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀ ⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀ ⠀⠤⢄⠀⠀⠀⠀⠀⠀⠀⠀⡼⠀⣧⠀⢿⢠⣤⣤⣬⣥⠀⠁⠀⠀⠛⢀⡒⠀⠀⠀⠘⡆⡆⠀⠀⠀⡇⠀⠀⠇⠀⠀⠀⠀⠀⠀⠀⠀⠀
⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀ ⠀⠀ ⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀ ⠀⢵⡀⠀⠀⠀⠀⠀⡰⠀⢠⠃⠱⣼⡀⣀⡀⠀⠀⠀⠀⠀⠀⠀⠈⠛⠳⠶⠶⠆⡸⢀⡀⣀⢰⠀⠀⢸ ⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀
⠀⠀⠀⠀ ⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀ ⣀⣀⣀⠄⠀⠉⠁⠀⠀⢠⠃⢀⠎⠀⠀⣼⠋⠉⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠴⠢⢄⡔⣕⡍⠣⣱⢸⠀⠀⢷⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀
⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀ ⠀⠀⠀⠀⠀ ⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀ ⠀⠀⠀⡰⠃⢀⠎⠀⠀⡜⡨⢢⡀⠀⠀⠀⠐⣄⠀⠀⣠⠀⠀⠀⠐⢛⠽⠗⠁⠀⠁⠊⠀⡜⠸⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀
⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀ ⠀ ⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀ ⠀⠀⠀⠀⠀ ⢀⠔⣁⡴⠃⠀⡠⡪⠊⣠⣾⣟⣷⡦⠤⣀⡈⠁⠉⢀⣀⡠⢔⠊⠁⠀⠀⠀⠀⢀⡤⡗⢀⠇⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀
⠀ ⠀⠀⠀⠀ ⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀ ⠀⠀⢀⣠⠴⢑⡨⠊⡀⠤⠚⢉⣴⣾⣿⡿⣾⣿⡇⠀⠹⣻⠛⠉⠉⢀⠠⠺⠀⠀⡀⢄⣴⣾⣧⣞⠀⡜⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀
⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀ ⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀ ⠐⠒⣉⠠⠄⡂⠅⠊⠁⠀⠀⣴⣿⣿⣿⣿⣻⣿⣿⡇⠀⠀⢠⣷⣮⡍⡠⠔⢉⡇⡠⠋⠁⠀⣿⣿⣿⣿⣄⠀⠀⠀⠀
Jan 23, 4:28 AM

Online
Sep 2016
21970
Reply to traed
@Zarutaku
You can ask random important questions too
https://www.wolframalpha.com/input?i=Average+bra+size

34DD for the US what....? Is it because so many fat people O.o
@traed Not sure, maybe they have different size systems. The US is well known to use their own standards instead of ISO
*kappa*
Jan 23, 5:10 AM
Offline
Jun 2021
2748
Science? Probably not.
It probably is over for programerbros however. Much cheaper to have the AI program barf out code and for the coder to inspect and correct mistakes.
Hot Blood saves lives.
Jan 24, 10:32 PM
lagom
Online
Jan 2009
107440
Google DeepMind CEO Demis Hassabis says AGI that is robust across all cognitive tasks and can invent its own hypotheses and conjectures about science is 3-5 years away https://www.reddit.com/r/singularity/comments/1i8w57v/google_deepmind_ceo_demis_hassabis_says_agi_that/

More topics from this board

» The Rise Of AI? ( 1 2 )

DigiCat - Oct 2

52 by AmityBlight »»
42 seconds ago

» am I a bitch for this?

Commit_Crime - Oct 8

24 by ISeeLifePeople »»
8 minutes ago

» Have you ever used counterfeit money to purchase something?

VabbingSips - 1 hour ago

3 by Zarutaku »»
56 minutes ago

» If AI robots get advanced enough to bang, would you?

Neroir - 4 hours ago

15 by deg »»
2 hours ago

» Silly ways of making money

traed - Yesterday

13 by LSSJ_Chloe »»
2 hours ago
It’s time to ditch the text file.
Keep track of your anime easily by creating your own list.
Sign Up Login