Page 1
Page 1
Started By
Message

Google's AI just solved 84% of the International Math Olympiad (IMO) problems from 2000-24

Posted on 2/7/25 at 12:36 pm
Posted by rickgrimes
Member since Jan 2011
4255 posts
Posted on 2/7/25 at 12:36 pm
Posted by Sun God
Member since Jul 2009
46738 posts
Posted on 2/7/25 at 12:38 pm to
What did it think of our playbook
Posted by TigerMonkey
Beach
Member since Jul 2005
7271 posts
Posted on 2/7/25 at 12:38 pm to
So does this mean we are winning the CWS this year?
Posted by kjp811
Denver, CO
Member since Apr 2017
964 posts
Posted on 2/7/25 at 12:39 pm to
Only 84%?
Posted by Zephyrius
Wharton, La.
Member since Dec 2004
8578 posts
Posted on 2/7/25 at 12:44 pm to
No wonder why we are Wide Receiver U
Posted by Gene Heinous
Houbuston
Member since Sep 2021
507 posts
Posted on 2/7/25 at 1:22 pm to
When it had previously calculated that the population with maximum efficiency for existing on the Earth was around 750,000; it now also calculates that billionaires are more worthy to be in that number than non-billionaires.
Posted by lsufanva
sandston virginia
Member since Aug 2009
13082 posts
Posted on 2/7/25 at 1:22 pm to
quote:

Only 84%?


Does seem low.
If AI and the most brilliant minds in the world can't solve these problems, how do we know what the correct answers are?
Posted by Fun Bunch
New Orleans
Member since May 2008
123604 posts
Posted on 2/7/25 at 1:34 pm to
Shouldn’t it be able to solve all of them?
Posted by S
RIP Wayde
Member since Jan 2007
164278 posts
Posted on 2/7/25 at 1:35 pm to
tEgg in tMud
Posted by drizztiger
Deal With it!
Member since Mar 2007
42761 posts
Posted on 2/7/25 at 1:41 pm to
You know when someone gets this lost on where to post, they are almost assuredly a Poli talk poster.

…and a quick post history shows yet again that is correct. It’s more than 84% of the time.
Posted by GumboPot
Member since Mar 2009
132652 posts
Posted on 2/7/25 at 1:43 pm to
I tested an advanced algebra problem on Gemini, ChatGPT and DeepSeek:

"Solve the real roots for k in the equation (1+k)^10 + 4(1+k)^5 - 5 = 0"

Gemini and DeepSeek nailed it. ChatGPT could only find one root and I had to coax it along to get to the other real root.

FYI: k = 0 and k = (-5)^(1/5) - 1 are the real roots.

ETA: wait, why is this on the rant?
This post was edited on 2/7/25 at 1:44 pm
Posted by baytiger11
Member since Jul 2020
1924 posts
Posted on 2/7/25 at 1:45 pm to
quote:

Only 84%
Ha, AI got a B-
Posted by JerryTheKingBawler
South of Memphis
Member since Jan 2023
4846 posts
Posted on 2/7/25 at 1:48 pm to
Les Miles has lost control of AI
Posted by TheDrunkenTigah
Baton Rouge
Member since Aug 2011
17855 posts
Posted on 2/7/25 at 1:49 pm to
Maybe I’m wrong but it doesn’t seem all that impressive that what is essentially a sophisticated calculator is decent at math.
Posted by Kato
Sec 102
Member since Nov 2006
2926 posts
Posted on 2/7/25 at 1:49 pm to
quote:

84%

Posted by Hotgin
Kazakhstan
Member since Jan 2014
1443 posts
Posted on 2/7/25 at 1:53 pm to
Wow a calculator worked!!!
Posted by Woolfpack
Member since Jun 2021
982 posts
Posted on 2/7/25 at 2:01 pm to
Bonus points if it said “I dont know” for the questions it missed.
Posted by Master of Sinanju
Member since Feb 2012
11638 posts
Posted on 2/7/25 at 2:06 pm to
Computers are good at math. News at 11.
Posted by Green Chili Tiger
Lurking the Tin Foil Hat Board
Member since Jul 2009
49898 posts
Posted on 2/7/25 at 2:28 pm to
Do the community notes show up after you guys embed the tweet or are you guys just saying frick it and ignoring it?
Posted by POTUS2024
Member since Nov 2022
20943 posts
Posted on 2/7/25 at 2:35 pm to
Garbage
first pageprev pagePage 1 of 1Next pagelast page
refresh

Back to top
logoFollow TigerDroppings for LSU Football News
Follow us on X, Facebook and Instagram to get the latest updates on LSU Football and Recruiting.

FacebookXInstagram