Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

The language brain matters more for programming than the math brain? (2020)

massivesci.com

Post reply on HN
Print view View original

161–170 of 465 posts

Re: The language brain matters more for programming than the math brain? (2020)

#161

Post by Qem »

Jensson wrote: post #120
codr7 wrote: post #94

Earlier quoted context omitted.

Low level requires more discipline, similar to math, that's about it.

Math doesn't require discipline though, many learn and do math without much discipline.

It surely helps. Take for example History, if I space out for whatever reason in the class and botch my examination on the Aztecs, that doesn't preclude me doing great in the classes about the Roman empire. History is like a sparse suburb, where a single house catching fire does not necessarily propagate and destroy the others. On the other side Math its like a skyscrapper, one floor build upon the other. If someone living in a lower floor botches a home renovation in his unit and damages a pillar, it risks taking down the whole building. In math if you botch a prerequisite, you'll ending messing your learning of subsequent concepts 95% of the time. So it is more unforgiving to people with problems like ADHD, that struggle to keep unbroken focus and discipline.

Re: The language brain matters more for programming than the math brain? (2020)

#162

Post by Jensson »

deeThrow94 wrote: post #41
karmakaze wrote: post #38

Earlier quoted context omitted.

> Recursion is also fundamentally linguistic You sure about that? How about inductive proofs? I would just say that language is more familiar to most. Mathematics are also languages, but more formal and foreign to most.

How do you communicate an inductive proof without language? Even formal symbolic logic is fundamentally linguistic.

> How do you communicate an inductive proof without language?

With that argument everything is fundamentally linguistic since everything is communicated using a language.

Can you come up with a more reasonable argument?

Re: The language brain matters more for programming than the math brain? (2020)

#163

Post by NickM »

77pt77 wrote: post #82
divbzero wrote: post #34

So maybe it makes sense that LLMs do okay at programming even though they lack the ability to reason?

They don't do okay. Quite on the contrary. My experience is that they spit out reasonably looking solutions but then they don't even parse/compile. They are OK to create small spinets of code and completion. Anything past that they suck. It's actually hilarious that AI "solved" bullshiting and and artistic fields much better and faster than say reasoning fields like math or programming. It's the supreme irony. Even 5…

I disagree that the current generation of AI has "solved" artistic fields any more than it's solved math or programming.

Just as an LLM may be good at spitting out code that looks plausible but fails to work, diffusion models are good at spitting out art that looks shiny but is lacking in any real creativity or artistic expression.

Re: The language brain matters more for programming than the math brain? (2020)

#164

Post by necovek »

godelski wrote: post #113

It think this is silly on multiple accounts. I'll claim that there's not real thing such as a "language brain" or "math brain." I'll also claim that most people don't know what math is, and that their evidence supports a "math brain". Math isn't about calculations/computations, it is about patterns. You get to algebra and think "what are these letters doing in my math" but once you get further you think "what are the…

To be fair, the original study uses "numeracy", and the correlation numbers are almost exactly like the ones for "language aptitude".

At the same time, the study excluded "five participants ... due to attrition (not completing the training sessions), and one participant ... because he was an extreme outlier in learning rate (>3 sd away from the mean)." I mean, if you are to exclude 15% of your subjects without looking at their aptitude (maybe they didn't do it because it was too hard to pass the training tests to move to the next lesson, yet their language aptitude is high?), with only 36 subjects of which 21 are female (it's obvious programming is male dominated, so they only had 15 males: maybe it doesn't matter, but maybe it does), how can you claim any statistical significance with such small numbers?

Re: The language brain matters more for programming than the math brain? (2020)

#165

Post by apeescape »

godelski wrote: post #113

It think this is silly on multiple accounts. I'll claim that there's not real thing such as a "language brain" or "math brain." I'll also claim that most people don't know what math is, and that their evidence supports a "math brain". Math isn't about calculations/computations, it is about patterns. You get to algebra and think "what are these letters doing in my math" but once you get further you think "what are the…

Finland tried teaching maths to children using Group Theory in the 70s [1], but the results weren't that good; it proved to be too abstract for young kids.

Ultimately, I believe basic algebra and geometry are the most important takeaways from math classes for most people.

[1]: https://www.hs.fi/tiede/art-2000004823594.html (sorry, it's in Finnish and behind a paywall)

Re: The language brain matters more for programming than the math brain? (2020)

#166

Post by godelski »

NickM wrote: post #158
godelski wrote: post #113

It think this is silly on multiple accounts. I'll claim that there's not real thing such as a "language brain" or "math brain." I'll also claim that most people don't know what math is, and that their evidence supports a "math brain". Math isn't about calculations/computations, it is about patterns. You get to algebra and think "what are these letters doing in my math" but once you get further you think "what are the…

I think you might be missing the point of the article: the study being cited isn't trying to establish the existence of a "language brain" or a "math brain", that's just the way the headline editorialized it to help people understand the conclusions. The conclusion of the study was that linguistic aptitude seemed to be more correlated with programming aptitude than mathematical aptitude, which seems fairly interestin…

I understood it.

  > The conclusion of the study was that linguistic aptitude seemed to be more correlated with programming aptitude than mathematical aptitude
And this is what I'm pushing back against and where I think you've misinterpreted.

  > They found that how well students learned Python was mostly explained by general cognitive abilities (problem solving and working memory), while how quickly they learned was explained by both general cognitive skills and language aptitude.
I made the claim that these are in fact math skills, but most people confuse with arithmetic. Math is a language. It is a language we created to help with abstraction. Code is math. There's no question about this. Go look into lambda calculus and the Church-Turing Thesis. There is much more in this direction too. And of course, we should have a clear connection to connect it all if you're able to see some abstraction.

Re: The language brain matters more for programming than the math brain? (2020)

#167

Post by HarHarVeryFunny »

zahlman wrote: post #62
armchairhacker wrote: post #40

Good code doesn’t just solve a problem, it solves it in a way that’s readable and modular. I think the problem-solving part of coding requires math skills, while the organization part requires writing skills. The organization part affects the problem-solving part, because if you write messy code (that you can’t reread once you forget or extend without rewriting) you’ll quickly get overwhelmed. Writing large math proo…

> Whereas programming has more “small steps”: even someone who’s not smart (but has grit) can write an impressive program, if they write one component at a time and there aren’t too many components that rely on each other. ... In my experience, learning to write one component at a time (and try the code, and make sure it works before proceeding) is itself a skill that many struggle to develop. Similarly for avoiding…

I guess it depends on what is meant by "everyone can learn to program". It's a bit like saying everyone can learn to write, or to do math, or play tennis, etc.

I'm sure everyone is capable of learning some basic level of programming, just as they are able to learn a basic (high school) level of any subject. However, not everyone is going to have the aptitude to take that to an advanced professional level, however hard they try. We're not all cut out to be artists, or writers, or doctors, or scientists, or developers, etc.

Re: The language brain matters more for programming than the math brain? (2020)

#168

Post by godelski »

necovek wrote: post #164
godelski wrote: post #113

It think this is silly on multiple accounts. I'll claim that there's not real thing such as a "language brain" or "math brain." I'll also claim that most people don't know what math is, and that their evidence supports a "math brain". Math isn't about calculations/computations, it is about patterns. You get to algebra and think "what are these letters doing in my math" but once you get further you think "what are the…

To be fair, the original study uses "numeracy", and the correlation numbers are almost exactly like the ones for "language aptitude". At the same time, the study excluded "five participants ... due to attrition (not completing the training sessions), and one participant ... because he was an extreme outlier in learning rate (>3 sd away from the mean)." I mean, if you are to exclude 15% of your subjects without lookin…

  > the original study uses "numeracy"
That's fair, although I don't think it changes my response. And the article still really leads to the wrong conclusions. You want to teach children abstraction and reasoning? You teach them math. Not numeracy, math.

Re: The language brain matters more for programming than the math brain? (2020)

#169

Post by HideousKojima »

zahlman wrote: post #62
armchairhacker wrote: post #40

Good code doesn’t just solve a problem, it solves it in a way that’s readable and modular. I think the problem-solving part of coding requires math skills, while the organization part requires writing skills. The organization part affects the problem-solving part, because if you write messy code (that you can’t reread once you forget or extend without rewriting) you’ll quickly get overwhelmed. Writing large math proo…

> Whereas programming has more “small steps”: even someone who’s not smart (but has grit) can write an impressive program, if they write one component at a time and there aren’t too many components that rely on each other. ... In my experience, learning to write one component at a time (and try the code, and make sure it works before proceeding) is itself a skill that many struggle to develop. Similarly for avoiding…

>"absolutely everyone can learn to program"

Personally I've always considered a solid grasp of algebra to be the minimum bar for being able to program, at least for anything that isn't utterly trivial. Being able to take a word problem and turn it into a system of equations and solve it is a pretty close analog to being able to take some sort of problem or business requirement and turn it into code.

And the sad truth is that a huge percentage of the population struggle with just arithmetic, let alone algebra.

Re: The language brain matters more for programming than the math brain? (2020)

#170

Post by 77pt77 »

NickM wrote: post #163
77pt77 wrote: post #82

Earlier quoted context omitted.

They don't do okay. Quite on the contrary. My experience is that they spit out reasonably looking solutions but then they don't even parse/compile. They are OK to create small spinets of code and completion. Anything past that they suck. It's actually hilarious that AI "solved" bullshiting and and artistic fields much better and faster than say reasoning fields like math or programming. It's the supreme irony. Even 5…

I disagree that the current generation of AI has "solved" artistic fields any more than it's solved math or programming. Just as an LLM may be good at spitting out code that looks plausible but fails to work, diffusion models are good at spitting out art that looks shiny but is lacking in any real creativity or artistic expression.

> "looks shiny but is lacking in any real creativity or artistic expression."

My experience with that is that artistic milieus now sometimes even explicitly admit that the difference is who created the art.

"Human that suffered and created something" => high quality art

"The exact same thing but by a machine" => soulless claptrap

It's not about the end result.

A lot could be written about this but it's completely socially unacceptable.

Whether an analogous thing will happen with beautiful mathematical proofs or physical theories remains to be seen. I for one am curious, but as far as art is concerned, in my view it's done.

Post reply on HN