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

LLMs Will Replace 8-Track Duplication Engineers

bbenchoff.github.io

Post reply on HN
Print view View original

11–14 of 14 posts

Re: LLMs Will Replace 8-Track Duplication Engineers

#11

Post by razorbeamz »

ithkuil wrote: post #3
razorbeamz wrote: post #2

LLMs are not the ideal tool for this job, because LLMs cannot do math or count.

But LLM can write code that can do math and count. Tool use, more broadly, has proven to be a very powerful way to let LLMs do what they're good at (handle the fuzzy and imprecise nuances of natural language, which includes the scooping of a lot of context) and delegate other things they're not good at to external tools, some of which if can write on the spot. If you think about it, we humans do that all the time too…

> But LLM can write code that can do math and count.

They cannot, however, execute that code. They can feed that code into an external program they've been given access to, but they can't execute it themselves.

Re: LLMs Will Replace 8-Track Duplication Engineers

#12

Post by razorbeamz »

vitally3643 wrote: post #6
razorbeamz wrote: post #2

LLMs are not the ideal tool for this job, because LLMs cannot do math or count.

Most human programmers are also fantastically bad at math.

True but irrelevant.

This 8-track duplication puzzle is a problem of math.

Re: LLMs Will Replace 8-Track Duplication Engineers

#13

Post by in-silico »

razorbeamz wrote: post #10
in-silico wrote: post #7

Earlier quoted context omitted.

> LLMs cannot do math This is plainly not true anymore

No, they fundamentally cannot do math. They are next token predictors, not calculators.

Why can't a next token predictor do math? Humans aren't calculators either, but we can do math.

If you want proof just look at the benchmarks. Modern frontier models can get basically perfect accuracy on American Invitational Mathematics Examination tests: https://matharena.ai/?comp=aime--aime_2026

If you want an explanation of how they do math, we've found geometric calculators inside their neural networks: https://www.goodfire.ai/research/a-geometric-calculator#

Re: LLMs Will Replace 8-Track Duplication Engineers

#14

Post by ithkuil »

razorbeamz wrote: post #11
ithkuil wrote: post #3

Earlier quoted context omitted.

But LLM can write code that can do math and count. Tool use, more broadly, has proven to be a very powerful way to let LLMs do what they're good at (handle the fuzzy and imprecise nuances of natural language, which includes the scooping of a lot of context) and delegate other things they're not good at to external tools, some of which if can write on the spot. If you think about it, we humans do that all the time too…

> But LLM can write code that can do math and count. They cannot, however, execute that code. They can feed that code into an external program they've been given access to, but they can't execute it themselves.

You presumably have no problem moving around in a car that you only control indirectly via a steering wheel, an accelerator and a brake pedal without ever actually powering the wheels
Post reply on HN