The data doesn't well support the claim that FP is best. Elixir tops the table at 97.5%, but C# (88.4%) is OOP and scores almost identically to Racket (88.9%), and Ruby (81.0%) and Java (80.9%) both outscore Scala (78.4%), which is explicitly functional. If FP were the driver, Scala should beat those languages, but it doesn't.
It's tempting to argue that a more constrained language helps, but Rust (62.8%) vs Elixir (97.5%) is an interesting data point here. Both are highly constrained, but in different directions. Elixir's constraints narrow the solution space because you can't mutate, you can't use loops, and you must pattern match, so every constraint eliminates options and funnels you toward fewer valid solutions that the LLM has to search through. Rust adds another constraint that must independently be satisfied on top of solving the actual problem, where the borrow checker doesn't eliminate approaches but adds a second axis of correctness the LLM has to get right simultaneously.
Overall, it seems like languages with strong conventions and ecosystems that narrow the solution space beat languages where there's a thousand ways to do something. Elixir has one build tool, one formatter, one way to do things. C#, Kotlin, and Java have strong ceremony and convention that effectively narrow how you write a program. Meanwhile JS, Python, PHP, and Perl offer endless choices, fragmented ecosystems, and rapidly shifting idioms, and they cluster at the bottom of the table.
Sharlin 41 minutes ago [-]
If we assume that the amount of training data matters at least a bit (which is s very reasonsble asdumption), I wouldn’t immediately discard the functional hypothesis. Scala’s score is almost equal to Java’s even though there’s probably something like two orders of magnitude less Scala than Java code in the wild. Similarly with C# and Racket.
nr378 20 minutes ago [-]
Yep I think you can reasonably argue that immutability + strong conventions are the most important dimensions (as opposed to FP vs. OOP, as much as I like FP and dislike OOP):
I agree with you, but, from the article: "The amount of training data doesn’t matter as much as we thought. Functional paradigms transfer well"
Anyway, I tend to think you are right, and the article is wrong in that sentence. (Or I misinterpreted something?)
I think both the quantity and quality of that has a big influence in the results.
ashirviskas 37 minutes ago [-]
What if it is the quality of data? Internet is full of terrible python/js, but probably not Elixir.
deflator 29 minutes ago [-]
Seems plausible. I used to refer to StackOverflow before LLMs and a good amount of the examples there were flawed code presented as working. If the LLM had less junk in its training then it might benefit even though the volume of training on that language is lower.
arbirk 1 hours ago [-]
1) AutoCodeBranch does not present similar problems, so you can't compare the Upper Bound metric across languages. Actually the Elixir tests are transmuted from other languages.
2) The Tesla section is interesting. I'm not saying that you are wrong, just that their methods have not produced the promised results yet
3) Wireless humanoid robots is a bad platform because we don't have the hardware to support them. Both battery density and compute efficiency is too low currently to support freestanding robots. Rip Roomba - long live its legacy
jna_sh 2 hours ago [-]
> The amount of training data doesn’t matter as much as we thought.
Seems a huge assumption to me. From the data one could equally conclude that JavaScript and Python have lower code quality _because_ the quantity of training data, e.g. more code written by less experienced developers
zephen 1 hours ago [-]
Don't worry! It'll only get better, as the amount of ouroboros training data explodes.
Weren't we taught that recycling is good?
ChrisMarshallNY 1 hours ago [-]
It seems that this is really saying that Functional Programming is The Way.
Maybe so, but I find the right tool, for the right job, is important. Trying to use a screwdriver for every task, is going to result in bruises and cuts (and gouges in the work).
I've used a pretty "chimeric" approach, for most of my career, and the LLM that I use, doesn't seem to have an issue with that. Many of its responses, seem to take a similar approach.
le-mark 1 hours ago [-]
I don’t think it’s a surprise that statically typed, immutable functional languages are the easiest (most suitable?) translation target for LLMs. I am surprised Haskell wasn’t represented in the table. Ocaml and Standard ML would be interesting as well.
Is there a future in which functional languages finally see wide adoption due to this LLM suitability? I can’t say that I would object!
bronlund 1 hours ago [-]
Brilliant take! We have spent so much time getting stupid people to program, that we have actually made the machines dumber :D
WillAdams 2 hours ago [-]
If the tool chain needs to be structured to facilitate people writing programs and documentation and comments, why isn't Literate Programming more popular?
I found it interesting that Elixir scores so high, but I'm not sure whether I can agree with the cause.
Bolwin 3 hours ago [-]
That benchmark is useless for comparing languages because the tasks are not the same across languages
gostsamo 3 hours ago [-]
how can you argue with so many assertive sentences in the article? they leave no space for critical thinking.
ashirviskas 30 minutes ago [-]
I'll admit, my brain was DDoSed by the article and I thought that maybe posting it here will get us someone with more DDoS proof brain to dissect it.
austin-cheney 2 hours ago [-]
> I’m still seeing a decent number of people on Twitter complain occasionally that they’ve tried AI-driven coding workflows and the output is crap and they can move faster by themselves. There’s less of these people in the world of Opus 4.5 and Gemini 3 now, but they’re still there.
The article starts from a false premise: that AI assisted coding makes the code more understandable. This isn't the case. You either understand the code without AI or offload that reasoning onto the AI, at which point its not you that understands the code.
A person could argue AI writes original code more understandable at maintenance time than they could on their own. This is equally problematic for the same reason. If a person has a lesser understanding of the code at original authoring they will have a lesser understanding of the edge cases and challenges that went into the reasoning about that original code and its those thought challenges which inform the complexities of maintenance, not the simplicity of the base code.
As an analog its like being given a challenging game puzzle to solve. After realizing the game requires extended effort to reach the desired goal the person searches online for the puzzle solution. At the game's next level they encounter a more challenging puzzle, but they never solved the prior puzzle, and so cannot solve this puzzle. In effect all understanding is destroyed and they have become utterly reliant on spoon-fed solutions they cannot maintain themselves.
keybored 2 hours ago [-]
Against Flintstone Engineering.[1] That’s great.
I don’t know about the premises here. All of these articles are written to hammer two points.
- AI is the future/AI has been here since X months ago
- There are still people who don’t believe that—to me an unfathomable position as I have personally spent five gazillion tokens on
And the supposed topic of the article is incidental to that.
But if GenAI is the future I’ll take GenAI formal verification and code generation over mindless code generation, thank you very much.
This is so annoying. First it’s obviously AI generated so it’s hard to even read. But if we get past that it’s making all kinds of uncited claims. Did Grace Hopper envision the translation layer moving directly from English to machine code? I don’t know because I can’t trust the LLM to say, and the article does not include a citation — in a piece whose central claim is that AI shifts the burden from coding to verification.
stabbles 3 hours ago [-]
The TL;DR: code should be easy to audit, not easy to write for humans.
The rest is AI-fluff:
> This isn't about optimizing for humans. It's about infrastructure
> But the bottleneck was never creation. It was always verification.
> For software, the load-bearing interface isn't actually code. Code is implementation.
> It's not just the Elixir language design that's remarkable, it's the entire ecosystem.
> The 'hard' languages were never hard. They were just waiting for a mind that didn't need movies.
tyleo 2 hours ago [-]
To put it another way: this article isn’t about the AI fluff, it’s about the two sentences at the top the author wrote themselves. ;)
zeristor 2 hours ago [-]
Perhaps we need an AI to human transformer to remove the AI fluff?
zeristor 2 hours ago [-]
It really is AI fluff.
Are people starting to write and talk in this manner, I see so many YouTube videos where you can see a person reading an AI written text, its one thing if the AI wrote it, but another if the human wrote it in the style of an AI.
As someone pointed out to me the way an AI writes text can be changed, so it is less obvious, its just that people don't tend to realise that.
arrowsmith 2 hours ago [-]
It's even infecting the highest levels of government:
"X isn't A, it's (something opposite A)" I twitch involuntarily.
ModernMech 2 hours ago [-]
Someone had one of those AI videos on in the background and, I can’t explain it, the ordering of the words is like nails on a chalkboard to me. I’m starting to have a visceral physiological response to AI prose that makes it actually painful to listen to.
The video was a biography about some Olympian, and I could tell the prompt included some facts about her wanting to be a tap dancer as a kid, because the video kept going back to that fact constantly. Every few sentences it would reference “that kid who wanted to be a tap dancer”. By the 6th time it brought up she wanted to be a tap dancer I was ready to scream.
dist-epoch 3 hours ago [-]
Man you are bad at TL;DR;-ing, you completely left out the main point article makes comparing stateful/mutating object oriented programming that humans like and pure functional oriented programing that presumably according to author LLMs thrive in.
skywhopper 3 hours ago [-]
This article takes a very tiny, questionable bit of data and extrapolates a lot of iffy assertions.
In general I’m tired of the “humans need never, and should never look at the code” LLM triumphalism articles. Do these folks ever work with real systems, I wonder.
dist-epoch 2 hours ago [-]
I remember when "real programmers" were supposed to look at the assembly code generated by compilers because it was bloated, inefficient, and totally unsuitable to use in a real system.
Cue in "non-determinism" retort.
tgv 2 hours ago [-]
Hardware restrictions might have contributed to that. Anyway, analogs and metaphors do not prove what they sneakily try to imply. They might help thinking about a problem, but they leave out the actual argument, and in this case, the jump is substantial.
chrisrhoden 2 hours ago [-]
I think the problem is less determinism than predictability. Hashing algorithms are deterministic.
Will people start .gitignore-ing their src directories and only save prompts?
ModernMech 2 hours ago [-]
This article [1] would argue ”no”, because then you would be ridding yourself of a “repository of determinism”, which the prompts cannot replace.
You can build a system with non-deterministic properties but you need some sort of deterministic foundation to build working, usable systems. Non determinism from top to bottom is building on quicksand in a swamp.
There are multiple reasons why binary repository managers like Artifactory exist.
And, arguably, the primary reason that perforce is still popular in some domains is that it works well with both large and opaque objects.
Once you've decided to version your src directory, whether a priori or because once bitten, twice shy, the next question is:
Is git the correct version control for this? Or are all the changes so large that git's advantages in merging and manually diffing things become irrelevant?
That you anticipated a retort isn’t enough. You also have to refute it.
Yeah compilers are deterministic and LLMs are not. The response to that?
The answer could very well be something like what’s in TFA namely formal verification. But an answer here is needed.
papa0101 2 hours ago [-]
Surprised to see Go scoring so low
Chris2048 2 hours ago [-]
> We built objects with identity and state because that’s how we experience reality
I mean, we called them objects, but coupling related state (and functions) together seem an objectively (object-ively) way to group data, it's literally just dict-based organisation.
muskstinks 2 hours ago [-]
[dead]
Rendered at 13:21:44 GMT+0000 (Coordinated Universal Time) with Vercel.
It's tempting to argue that a more constrained language helps, but Rust (62.8%) vs Elixir (97.5%) is an interesting data point here. Both are highly constrained, but in different directions. Elixir's constraints narrow the solution space because you can't mutate, you can't use loops, and you must pattern match, so every constraint eliminates options and funnels you toward fewer valid solutions that the LLM has to search through. Rust adds another constraint that must independently be satisfied on top of solving the actual problem, where the borrow checker doesn't eliminate approaches but adds a second axis of correctness the LLM has to get right simultaneously.
Overall, it seems like languages with strong conventions and ecosystems that narrow the solution space beat languages where there's a thousand ways to do something. Elixir has one build tool, one formatter, one way to do things. C#, Kotlin, and Java have strong ceremony and convention that effectively narrow how you write a program. Meanwhile JS, Python, PHP, and Perl offer endless choices, fragmented ecosystems, and rapidly shifting idioms, and they cluster at the bottom of the table.
Immutable by convention + Strong conventions: 91.3% - Elixir 97.5%, Kotlin 90.5%, Racket 88.9%, C# 88.4%
Immutable by convention + Fragmented: 78.4% - Scala 78.4% (n=1)
Mutable + Strong conventions: 77.5% - Ruby 81.0%, Swift 78.5%, Julia 78.5%, Dart 78.0%, Go 71.7%
Mutable + Fragmented: 67.9% - Java 80.9%, R 75.8%, C++ 75.8%, Shell 72.9%, Python 65.3%, Perl 64.5%, TS 61.3%, JS 60.9%, PHP 53.8%
(my grouping is somewhat subjective)
Anyway, I tend to think you are right, and the article is wrong in that sentence. (Or I misinterpreted something?)
I think both the quantity and quality of that has a big influence in the results.
2) The Tesla section is interesting. I'm not saying that you are wrong, just that their methods have not produced the promised results yet
3) Wireless humanoid robots is a bad platform because we don't have the hardware to support them. Both battery density and compute efficiency is too low currently to support freestanding robots. Rip Roomba - long live its legacy
Seems a huge assumption to me. From the data one could equally conclude that JavaScript and Python have lower code quality _because_ the quantity of training data, e.g. more code written by less experienced developers
Weren't we taught that recycling is good?
Maybe so, but I find the right tool, for the right job, is important. Trying to use a screwdriver for every task, is going to result in bruises and cuts (and gouges in the work).
I've used a pretty "chimeric" approach, for most of my career, and the LLM that I use, doesn't seem to have an issue with that. Many of its responses, seem to take a similar approach.
Is there a future in which functional languages finally see wide adoption due to this LLM suitability? I can’t say that I would object!
http://literateprogramming.com/
Is LLM what will finally push LP into mainstream acceptance?
https://news.ycombinator.com/item?id=47300747
The article starts from a false premise: that AI assisted coding makes the code more understandable. This isn't the case. You either understand the code without AI or offload that reasoning onto the AI, at which point its not you that understands the code.
A person could argue AI writes original code more understandable at maintenance time than they could on their own. This is equally problematic for the same reason. If a person has a lesser understanding of the code at original authoring they will have a lesser understanding of the edge cases and challenges that went into the reasoning about that original code and its those thought challenges which inform the complexities of maintenance, not the simplicity of the base code.
As an analog its like being given a challenging game puzzle to solve. After realizing the game requires extended effort to reach the desired goal the person searches online for the puzzle solution. At the game's next level they encounter a more challenging puzzle, but they never solved the prior puzzle, and so cannot solve this puzzle. In effect all understanding is destroyed and they have become utterly reliant on spoon-fed solutions they cannot maintain themselves.
I don’t know about the premises here. All of these articles are written to hammer two points.
- AI is the future/AI has been here since X months ago
- There are still people who don’t believe that—to me an unfathomable position as I have personally spent five gazillion tokens on
And the supposed topic of the article is incidental to that.
But if GenAI is the future I’ll take GenAI formal verification and code generation over mindless code generation, thank you very much.
[1] https://news.ycombinator.com/item?id=47358696
The rest is AI-fluff:
> This isn't about optimizing for humans. It's about infrastructure
> But the bottleneck was never creation. It was always verification.
> For software, the load-bearing interface isn't actually code. Code is implementation.
> It's not just the Elixir language design that's remarkable, it's the entire ecosystem.
> The 'hard' languages were never hard. They were just waiting for a mind that didn't need movies.
Are people starting to write and talk in this manner, I see so many YouTube videos where you can see a person reading an AI written text, its one thing if the AI wrote it, but another if the human wrote it in the style of an AI.
As someone pointed out to me the way an AI writes text can be changed, so it is less obvious, its just that people don't tend to realise that.
https://www.pimlicojournal.co.uk/p/mps-are-almost-certainly-...
"X isn't A, it's (something opposite A)" I twitch involuntarily.
The video was a biography about some Olympian, and I could tell the prompt included some facts about her wanting to be a tap dancer as a kid, because the video kept going back to that fact constantly. Every few sentences it would reference “that kid who wanted to be a tap dancer”. By the 6th time it brought up she wanted to be a tap dancer I was ready to scream.
In general I’m tired of the “humans need never, and should never look at the code” LLM triumphalism articles. Do these folks ever work with real systems, I wonder.
Cue in "non-determinism" retort.
Will people start .gitignore-ing their src directories and only save prompts?
You can build a system with non-deterministic properties but you need some sort of deterministic foundation to build working, usable systems. Non determinism from top to bottom is building on quicksand in a swamp.
[1] https://www.oreilly.com/radar/can-language-models-replace-co...
And, arguably, the primary reason that perforce is still popular in some domains is that it works well with both large and opaque objects.
Once you've decided to version your src directory, whether a priori or because once bitten, twice shy, the next question is:
Is git the correct version control for this? Or are all the changes so large that git's advantages in merging and manually diffing things become irrelevant?
Yeah compilers are deterministic and LLMs are not. The response to that?
The answer could very well be something like what’s in TFA namely formal verification. But an answer here is needed.
I mean, we called them objects, but coupling related state (and functions) together seem an objectively (object-ively) way to group data, it's literally just dict-based organisation.