this post was submitted on 15 Nov 2024
425 points (94.2% liked)

Programmer Humor

19804 readers
345 users here now

Welcome to Programmer Humor!

This is a place where you can post jokes, memes, humor, etc. related to programming!

For sharing awful code theres also Programming Horror.

Rules

founded 2 years ago
MODERATORS
 

Probably should've just asked Wolfram Alpha

(page 2) 20 comments
sorted by: hot top controversial new old
[–] [email protected] 1 points 3 weeks ago* (last edited 3 weeks ago) (1 children)

It is correct. Half is 3/6 a third is 2/6. So a half is one third larger than 1/3

[–] Mercuri 1 points 3 weeks ago

You dropped this: /s

[–] givesomefucks 1 points 1 month ago* (last edited 1 month ago) (1 children)

Divide 50% by 3 and get 16.666666....

Add that to the 50%, you gets 66.6666...%

Which is the same as dividing a whole into 3 and adding two together.

It's trying to get there, but it seems to think you're talking about starting with 50% and increasing that by 1/3 of the existing half.

It interpreted what you typed as something like:

X=50

Y=x/3

Solve for:

X+Y

And gave you 66.666...%

But you could also just get a calculator and type 50 - 33.33333

I don't know why anyone would Google this question in a search engine, AI is dumb enough without trying to phrase things in a way that you think it says something stupid.

Like, did you legitimately Google it that way in an attempt find out the difference between half and a third?

[–] Mercuri 1 points 1 month ago (2 children)

It’s trying to get there, but it seems to think you’re talking about starting with 50% and increasing that by 1/3 of the existing half.

Yeah, it's definitely trying something. But it's also showing the problems with LLM.

Like, did you legitimately Google it that way in an attempt find out the difference between half and a third?

I remember being able to put basic math problems into Google and it spit out the correct answer. But you're right, that's entirely on me for expecting Google to work right after all the enshitification. For the record, Wolfram Alpha gave the correct answer.

[–] [email protected] 1 points 1 month ago

LLMs predict statistical likelihoods of words coming next. Them not being able to do math, unless explicitly given extra instructions to identify and do those problems, has been reported before, IIRC.

[–] givesomefucks 1 points 1 month ago

I remember being able to put basic math problems into Google and it spit out the correct answer

Pretty sure you can still do that if you don't type it out as weird as possible...

But I haven't tried to Google something in over a year.

Have fun tho I guess.

load more comments
view more: ‹ prev next ›