Hacker News Reader

Comment by culi | Hacker News Reader

My friend works with people in their 20s. She recently brought up her struggles to do the math in her head for when to clock in/out for their lunches (30 minutes after an arbitrary time). The young coworker's response was "Oh I just put it into ChatGPT"

The kids are using ChatGPT for simple maths...

View on HN

That'll lead to interesting results. I used a couple of LLM's for my blood bowl statistics, and they get rather simple math wrong. Which makes sense, they aren't build for math after all. It's wild how wrong they can get results though, I'd add the same prompt to 6 different AI's and they'd all get it wrong in 6 different ways.

On a side note, the most hilarious part of it was when I asked gemini to do something for me in Google Sheets and it kept refering to it as Excel. Even after I corrected it.

Are you sure the coworker wasn't joking? Because if somebody confessed to me they struggle to add half an hour to a time point, my first reaction would definitely be to laugh it off.

This is not at all the only instance I've heard of ridiculous levels of reliance on LLMs from her young coworkers (yes, multiple stories from multiple persons). This is just the most recent and most prominent in my mind.

I've been using a Python prompt or the browser URL bar for simple maths for over a decade. I don't see much added value in doing arithmetic manually, humans really suck at it.

You can’t add 30 minutes in your head?

LLMs are notoriously unreliable at math but even more than that it's about using the appropriate tool for the job. When you Google something, google is smart enough to give you a simple calculator. A simple LLM query like this uses about as much electricity as running a lightbulb for 15 minutes

It's easy to miss the value in something you don't do. I do fermi estimates in my head all the time and it would be exhausting to constantly pull out my phone to calculate things, to the point that I would stop attempting it as much as I do.

Humans don't suck at arithmetic.

Anecdata: Most cashiers used to be able to give correct change at checkout very quickly; only a few would type it into the register to have it do the math. Nowadays, with so many people using cards etc., many of them freeze up and struggle with basic change-making.

It's just a matter of keeping in practice and not letting your skills atrophy.

Eh. I have a math degree. Aced all the advanced maths. Was the only one to get an A in Diff Eq. I love math. I've never been able to do simple math in my head. I can't even remember the times tables half the time. Simple math isn't really problem solving.

It’s one type of problem solving.

That's absolutely valid but running a simple query to an LLM uses the amount of electricity as running a lightbulb for 15 minutes

It would've been faster to open up the calculator app and type in the numbers and get an instant response instead of opening up the ChatGPT app, typing in your question, waiting dozens of seconds, and getting a long response back.

People who major in mathematics are really good at mathematical abstraction and are _notorious_ for their inability to do basic arithmetic. To the degree that it's a stereotype with a strong grounding it reality.

In college we had a rule for splitting the check at a restaurant: the youngest non-math major had to do it. Not being a math major, I'm not sure what happened when the table was all math majors. It wasn't a frequent occurence; there was a strong likelihood of a physicist or an engineer being around.

But you could if you wanted to, probably.

It's ok this is just the next level of human evolution. We haven't needed to know how to do basic math since the calculator. Nowadays our AIs can read and write for us too. More obsolete skills. We can focus on higher level things now. No more focusing on sparks, we can focus on building something important. We don't have an attention span over 5 seconds anyway thanks to social media. If you don't get where I'm coming from you probably don't have ADHD but that's fine.

you need a spark to start a fire, if you offload everything to the LLM you won't understand the higher level things

Completely agree fwiw. My comment sarcastically paraphrased a few other AI slop lovers I've seen in this comment section.

Reply on news.ycombinator.com