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

Human coders are still better than LLMs

antirez.com

Post reply on HN
Print view View original

671–680 of 762 posts

Re: Human coders are still better than LLMs

#671

Post by skydhash »

vidarh wrote: post #666
skydhash wrote: post #643

Earlier quoted context omitted.

If you give it the rules to generate something, why can't it generate it? That's what something like Mockaroo[0] does. It's just more formal. That's pretty much what LLM training does, extracting patterns from a huge corpus of text. Then it goes one to generate according to the patterns. It can not generate a new pattern that is not a combination of the previous one. [0]: https://mockaroo.com/

> If you give it the rules to generate something, why can't it generate it? It can, but that does not mean that what is generate is not new, unless the rules in question constrains the set to the point where onely one outcome is possible. If I tell you that a novel has a minimum of 40,000 words, it does not mean that no novel is, well, novel (not sorry), just because I've given you rules to stay within. Any novel wil…

I was referring as new as some orthogonal dimension in the same space. If we're referring to your definition, any slight changes in the parameters results in something new. I was arguing more about if the model knows about axes x and y, then it's output is constrained to a plane unless you add z. But more often than not it's output will be a cylinder (extruded from a circle in the x,y plane) instead of a sphere.

The same thing goes for image generation. Every picture is new, but it's a combination of the pictures it founds. It does not learn about things like perspectives, values, forms, anatomy,... the way an artist does which are the proper dimensions of drawing.

> that humans somehow exceed the Turing computable

Already done by Gödel's incompleteness theorems[0] and the halting problem[1]. Meaning that we can do some stuff that no algorithm can do.

[0]: https://en.wikipedia.org/wiki/G%C3%B6del%27s_incompleteness_...

[1]: https://en.wikipedia.org/wiki/Halting_problem

Re: Human coders are still better than LLMs

#672

Post by the_af »

Aeolun wrote: post #517
daveguy wrote: post #354

Earlier quoted context omitted.

All. Of. Them. It's quite literally what they do because they are optimistic text generators. Not correct or accurate text generators.

Most optimistic text generators do not consider repeating the stuff that was already rejected a desireable path forward. It might be the only path forward they’re aware of though.

In some contexts I got ChatGPT to answer "I don't know" when I crafted a very specific prompt about not knowing being and acceptable and preferable answer to bullshitting. But it's hit and miss, and doesn't always work; it seems LLMs simply aren't trained to model admittance of ignorance, they almost always want to give a positive and confident answer.

Re: Human coders are still better than LLMs

#674

Post by addaon »

wvenable wrote: post #465
bigstrat2003 wrote: post #344

Earlier quoted context omitted.

> Perhaps you remember that language models were completely useless at coding some years ago, and now they can do quite a lot of things, even if they are not perfect. IMO, they're still useless today, with the only progress being that they can produce a more convincing facade of usefulness. I wouldn't call that very meaningful progress.

I don't know how someone can legitimately say that they're useless. Perfect, no. But useless, also no.

> I don't know how someone can legitimately say that they're useless.

Clearly, statistical models trained on this HN thread would output that sequence of tokens with high probability. Are you suggesting that a statement being probable in a text corpus is not a legitimate source of truth? Can you generalize that a little bit?

Re: Human coders are still better than LLMs

#675

Post by StillBored »

I think there is a common problem with a lot of these ML systems. The answers look perfectly correct to someone who isn't a domain expert. For example, I ask legal questions and it gives me fake case numbers I have no way to know are fake until I look them up. Same with the coding, I asked for a patch for a public project that has a custom !regex style match engine. It does an amazing job, cross referencing two different projects and hands me a very probable looking patch. I ask for a couple changes, one of which can't actually be done, but it creates some syntax that doesn't even compile because its using 'x' as a stand-in for the bits it doesn't have an answer for.

In the end, I had to go spend a couple hours reading the documentation to understand the matching engine, and the final patch didn't look anything at all like the LLM generated code. Which is what seems to happen all the time, it is wonderful for spewing the boilerplate, the actual problem solving portions its like talking to someone who simply doesn't understand the problem and keeps giving you what it has, rather than what you want.

OTOH, its fantastic for review/etc even though I tend to ignore many of the suggestions. Its like a grammar checker of old, it will point out you need a comma you missed, but half the time the suggestions are wrong.

Re: Human coders are still better than LLMs

#676

Post by albrewer »

Buttons840 wrote: post #388
Buttons840 wrote: post #379

Earlier quoted context omitted.

LLMs aren't my rubber duck, they're my wrong answer. You know that saying that the best way to get an answer online is to post a wrong answer? That's what LLMs do for me. I ask the LLM to do something simple but tedious, and then it does it spectacularly wrong, then I get pissed off enough that I have the rage-induced energy to do it myself.

I'm probably suffering undiagnosed ADHD, and will get stuck and spend minutes picking a function name and then writing a docstring. LLMs do help with this even if they get the code wrong, because I usually won't bother to fix their variables names or docstring unless needed. LLMs can reliably solve the problem of a blank-page.

> LLMs can reliably solve the problem of a blank-page.

This has been the biggest boost for me. The number of choices available when facing a blank page is staggering. Even a bad/wrong implementation helps collapse those possibilities into a countable few that take far less time to think about.

Re: Human coders are still better than LLMs

#677

Post by noslenwerdna »

vouaobrasil wrote: post #381
noslenwerdna wrote: post #315

Earlier quoted context omitted.

Everyone should do the tasks where they provide unique value. You could make the same arguments you just made for recorded music, automobiles, computers in general in fact.

Difference is though AI does it much faster and has much fewer central sources that provide the service. The speed and magnitude is important as well, just like a crash at 20km/h is different than a crash at 100km/h. And those other inventions WERE also harmful. Cars -> global warming.

My point is every invention has pros and cons, and tends to displace people who were very tied to the previous way.

Re: Human coders are still better than LLMs

#678

Post by Fernicia »

AstroBen wrote: post #458
Fernicia wrote: post #135

Earlier quoted context omitted.

If it automates 1/5th of your work, then what's unreasonable about thinking that your team could be 4 developers instead of 5?

If software costs 80% as much to write, what's unreasonable about thinking that more businesses would integrate more of it, hiring more developers?

I think both will happen. I was merely demonstrating that just because an AI can't replace the entirety of your work, doesn't mean it won't make you redundant.

Re: Human coders are still better than LLMs

#679

Post by skydhash »

vidarh wrote: post #668
skydhash wrote: post #652

Earlier quoted context omitted.

As long as you create the grammar, the language exists. Same if you edit a previous grammar. You're the one creating the language, not the model. It's just generating specific instance. If you tell someone that multiplying a number by 2 is adding the number to itself, then if this person knows addition, you can't be surprised if it tells you that 9*2 is 18. A small leap in discovery is when the person can extract the…

> As long as you create the grammar, the language exists. Yes, but it didn't exist during training . Nothing in the training data would provide pre-existing content for the model to produce from, so the output would necessarily be new. > But if you take the time to explain each concept.... Based on the argument you presented, nothing a human does is new, because it is all based on our pre-exististing learned rules of…

See my definition at : https://news.ycombinator.com/item?id=44137201

As mentioned there, I was arguing that without being prompted, there's no way that it can add something that is not a combination of the training data. And that combination does not act on the same terms that you would expect someone learning the same material would do.

In Linear regression, you can reduce a big amount of data to a small amount of factors. Every prediction would be a combination of those factors. According to your definition, those prediction will be new. For me what's new is when you retrospectively adds the input to the training data, find a different set of factors that gives you a bigger set of possible answers (generation) or narrows the definition of correct answers (reliability).

That is what people do when programming a computer. You goes from something that can do almost anything and you restrict it down to a few things (that you need). What LLM do is throwing the dice and what you get may or may not do what you want, and may not even be possible.

Re: Human coders are still better than LLMs

#680

Post by gitaarik »

vidarh wrote: post #647
gitaarik wrote: post #630

Earlier quoted context omitted.

But reward models are always curated by humans. If you generate a reward model with an LLM, it will contain hallucinations that need to be corrected by humans. But that is what a reward model is for. To correct the hallucinations of LLMs. So yeah theoretically you could generate reward models with LLMs, but they won't be any good, unless they are curated by other reward models that are ultimately curated by humans.

> But reward models are always curated by humans. There is no inherent reason why they need to be. > So yeah theoretically you could generate reward models with LLMs, but they won't be any good, unless they are curated by other reward models that are ultimately curated by humans. This reasoning is begging the question: The reasoning is true only if the conclusion is true. It's therefore a logically invalid argument.…

Sorry but I don't follow your logic. Are you claiming that reward models that aren't curated by humans perform as well as ones that are?

Then what is a reward model's function according to you?

Post reply on HN