José Valim & team have made such an incredible language and ecosystem. thank you for all the faithful work, especially the run up to 1.20 over this past year
Designs can't be easily split into a pure light or dark mode, though. Before light/dark modes were popularized, it was common to have some parts be "dark" and other parts be "light," depending on the design (for example, would the iOS 5 game center be "dark mode"?). Elixir's website uses a mix of light and dark to delineate sections.
One consequence of software recently supporting light and dark schemes is that UI designs end up committing to all white or all dark, rather than tastefully choosing a color scheme that best frames its content in context.
Yeah. You'll notice that there's a mix of light and dark. Some segments are light-on-dark, then it switches to dark-on-light. It appears to me like a "design trend" that's at odds with accessibility. https://www.apple.com does this too.
Why would it be at odds with accessibility? The text contrast is excellent in each section, and the light/dark segments clearly define the border between sections. Forced dark mode can be difficult to read for users with astigmatism, but the font size is large, and the actual documentation sites[0] do have a light/dark switch.
This, plus there is research that suggests dark mode does not help with retention of information.[1] I have started to slowly transition back to light mode for most of my work applications. (Email, Coding, etc.)
Dark mode feels correct, but it is hostile. Now blue light on the other hand... you should get filters for your monitor/glasses.
Reading the Erlang ProgProgrammers book by Joe Armstrong made me a better Ruby programmer as it changed my perspective on functional programming and abstractions.
I first reached for Elixir when Ruby couldn't handle large amounts of websocket messages. It really shines in high-concurrency contexts. I also love Phoenix LiveView and have a couple of side-projects running on it.
I think most will agree that it improved on Erlang.
For me as a long-term ruby user, though, elixir is
not quite as elegant as it could or should have been.
Even simple things such as "defmodule Xyz do" feels
weird to me.
I don't really have large monitors by today's standards, and the site looks nice enough but fully half of what I'm looking at is blank space. I don't remember what the old site looked like so don't know if that's really a change.
The majority of people who use LLMs today never even heard of ML though a non-trivial percentage have heard that modern AI is powered by LLM. You can’t forget what you never knew. Such is the evolution of language when a formerly niche technical concept crosses the chasm to mass awareness.
I'd argue there's a qualitative difference between using machine learning for specific data analysis tasks, and using a generic agentic AI system controlled by some corporate entity. The association of the term 'AI' with the latter is increasing.
Yes, but nozzlegear claims that even technically "intertwined" (presumably they mean "inclined") people don't know the connection between LLMs and the broader ML work that encompasses it. That's a pretty big claim, and would be rather shocking if true. ML and deep learning were heavily invested in and discussed through the 2010s (and earlier, but the hardware developments at the end of the 2000s enabled the ML boom of the 2010s), is our industry really so memory constrained (I know there's a shortage now, but still) that people don't know the connection between machine learning and LLMs?
> but nozzlegear claims that even technically "intertwined" (presumably they mean "inclined") people
Sorry, I meant the subjects (LLMs, ML, AI) are intertwined, not the people. But what I was getting at with my comment is that (IMO) most people see them as distinct things, even on HN where most know that LLMs use ML. As an analogy, it's like physics versus mathematics: separate subjects in most everyone's mind, and even separate academic departments, but physics is still math.
>I also worked on all of the copy myself, collecting feedback from core maintainers as I went. The new tagline was a suggestion from Theo which we iterated on. I did use LLMs as an assistant, but I did not ask it to generate the content.
>Might as well use LLMs for the whole thing next time, since we will be accused of doing so anyway! :D
There are multiple people working on the JIT within the last 5-6 years. The WhatsApp folks also contribute meaningfully.
I suspect once the Erlang/OTP team squeezes all performance in the JIT, they will look into optimizing across modules, which will probably open up many new possibilities, but it requires rethinking some runtime primitives.
A few years ago, I was working on an interpreter implemented in elixir for a domain specific language. It was a pretty basic metacircular interpreter. It relied heavily on function signature dispatch. When I tried breaking up the massive “interpret” function across modules, performance tanked. I got it all back by using some macro shenanigans, but understandably the team did not like this.
Knowing what I know now, I would’ve tried to push for a threaded interpreter to get rid of the runtime overhead of dispatching altogether. I don’t know if they’ve changed the architecture of that module much since I left :-)
1. The cross module optimizations I mentioned above
2. Have a WASM target for the runtime itself
3. Make it easier to ship single file executables with the whole VM
But they are really “nice-to-have”s. I have been a happy user for 15+ years!
i ran a quick experiment where instead of doing boxing the way its done in the beam currently, i used a different boxing (NaN strategy and there was a 10x speedup
To me, it seems one of the killer use cases for Elixir (/Erlang) is its distributed cluster capability. Does anyone have experience with that or case reports to share? I've used Elixir quite a bit professionally, but mostly as just a "nicer Rails" with horizontally scalable but otherwise independent Phoenix apps in your traditional Kubernetes setup, which seems to me to kind of missing out on its main purpose.
It's been a while, but I used to work at WhatsApp and we used Erlang distribution heavily. I understand the clusters have gotten really huge since I left.
It's super handy. There's no security barrier between nodes. It's a headache if your network is unreliable.
For a chat app, messaging someone becomes a series of steps:
a) look up if they're online (send a message to the presence database service)
b) if you got a process id back, that's the process connected to the user, so send it the message. The process could be on the same machine or not, but the sending api is the same. This is the special part: few other environments make arbitrary messaging between processes/threads/tasks/whathaveyou so pervasive.
c) if you don't get a process id back, the user is offline; send the message to the offline database.
Also interested in hearing about this! I built an elixir k8s control plane recently and kept expecting to reach for it but it never really made sense when it was controlling golang daemonsets.
My usecase is less independent though, that control plane is orchestrating like Lambda/fly.io style workloads on top of firecracker:
https://jomcgi.dev/ember
I've worked a little bit with distributed Elixir using `Horde.DynamicSupervisor` on Kubernetes. Apparently there's other options like 'swarm' and DynamicSupervisor [1]. It'd be great for clear analysis of the benefits these kinds of abstractions bring vs non-BEAM approaches.
Elixir/Erlang works very well in a semi-embedded environment where you need a higher level command and control component that behaves in a deterministic way and is pretty robust.
On the first syntax example: there’s something funny to me about using three pipe operator and four different functions to turn “hello world” into “Hello World”.
I appreciate Elixir but the problem is the job market/talent pool is tiny compared to other existing languages.
If you buy into the Elixir stack then you now have constraint you could've avoided entirely by avoiding it.
Also for devs there seems to be no premium offered for this talent pool scarcity. With LLMs I think language-specialists are redundant in a large scheme of things. ex) at one of my current remote jobs, I shipped an entire telecom infrastructure with barely knowing Elixir and we brought on contractors to audit the code and they found no issues.
> appreciate Elixir but the problem is the job market/talent pool is tiny compared to other existing languages.
> I shipped an entire telecom infrastructure with barely knowing Elixir and we brought on contractors to audit the code and they found no issues.
Erlang/Elixir experience is rare, because it's not widely used and the teams are small. It's not worth trying to hire for it. Hire for people who can figure it out on the go (amd are willing to give it a try).
You did it, hire other people who seem likely to be able to.
as a SWE this is not a good sign. it means the job market is slowly transitioning into temp work like economics. The value I got out of the Elixir contractors was immense since it not only proved that we can get a huge bulk of the work done without specialists and use them on demand for audit for a few months before AI this would've been not been possible.
normal market dynamics suggest scarcity demand premiums but this is not the case with software developers it seems.
a) did you pay your Elixir contractors more than you would pay a Java contractor for similar work?
but also...
b) scarcity isn't the only factor in price. Erlang/Elixir developers are scarce, but Erlang/Elixir jobs are also scarce. You need both demand and scarcity to raise prices. Also, it doesn't cost much to turn a willing, good developer into an Erlang/Elixir developer; substitute goods reduce the impact of scarcity.
also c) if you found contractors, but not employees, maybe you weren't willing to pay enough... So maybe the price is higher than you thought?
both contracts have over 10 years of experience with Elixir and one of them have written a widely used library. I think you are tad out of touch with the job market and with where agentic coding is right now.
Possibly, but I’m a senior software engineer that’s been writing Elixir for the last 5 years and has been experimenting with and using AI for the last 12-18 months.
Congrats on being one of the mythical developers that manages to get AI to write perfect code consistently!
i hired a biologist (for my pharma startup) and she produced feature ideas for our internal stack and was guiding claude to write idiomatic code with feedback from my reviews with no coding experience. realistically if you want to start an elixir company today you need one consciencious senior that likes code review and any number of juniors with minimal competency and sufficient curiosity.
Ugh, it looks like all the other LLM generated language webpages. It's formulaic at this point. I'd hoped a language like Elixir would be able to hire some people to do it.
The Software Mansion folks designed it and we actually iterated on the designs on Figma, having discussions as humans, and exploring alternatives. They were lovely to work with.
I also worked on all of the copy myself, collecting feedback from core maintainers as I went. The new tagline was a suggestion from Theo which we iterated on. I did use LLMs as an assistant, but I did not ask it to generate the content.
Might as well use LLMs for the whole thing next time, since we will be accused of doing so anyway! :D
Shame then that despite all that, they landed on the same design used by every "I asked an LLM to make me a language and a website this weekend here's what it spit out" project. I mean, I'm not saying it looks bad or is a bad result. Just it's very similar to other things that have put in much less effort.
It is funny (and perhaps a bit depressing) that LLMs were trained on our content and now, if we generate a similar structure as before, with the usual love and care, we will be criticized by it. Even when it does not "look bad or is a bad result".
It is indeed funny/sad to see the rancor created by our own inputs. You can't escape the AI police on HN right now. There will always be someone leaving a shallow complaint or accusation about the look or language, regardless of reality. Obviously slop exists, but the culture war has polarized some people enough that they're blinded, and anything resembling LLM output is a trigger. God help you if you use useful language like "load-bearing", because the police are definitely on their way.
FWIW, I don't think the site fits the LLM template. The scroll through the use cases is particularly nice.
And thanks for Elixir. I love it, and the agent + tidewave loop is a joy to use!
I’m tried to understand the motivation for this salty comment and the parent comment. I failed. Then I opened the user’s comment history and most of their comments are like this. ModernMech, please, keep in mind we’re all doing our best. Being passive aggressive on the internet is social pollution. No offense intended, I’m just hoping you reflect a bit next time before you post.
My motivation is I don't like the redesign and I wanted to say so. It looks like a lot of other sites out there. I'm sure that in the construction of the website got a lot of positive feedback along the way that caused such a result, but my feedback is that it looks like it could have been generated by an LLM. I didn't even say it looks bad, sorry if that's too harsh for the internet. You don't have to go on a quest to figure out my damage, it's not that deep.
Its pretty snappy/responsive for me at least so thats good. Normally LLM slop sites are pretty at first but sluggish as hell. So some level of skill went into this one.
Looks nice. But it would be more important
to clean up elixir itself. So many things
are unnecessary syntax-wise. At the least
elixir made working with erlang easier, so
they solved that part.
If you must default to dark mode that’s your choice but I’d love to see a light mode toggle somewhere prominent.
One consequence of software recently supporting light and dark schemes is that UI designs end up committing to all white or all dark, rather than tastefully choosing a color scheme that best frames its content in context.
[0]: https://elixir.hexdocs.pm/1.20.2/
Dark mode feels correct, but it is hostile. Now blue light on the other hand... you should get filters for your monitor/glasses.
[1] https://www.nngroup.com/articles/dark-mode/
Minor typo in the Erlang card:
“Elixir also excels at IoT, distributed systems, and everything the Erlang is renowned for”
should probably be “everything the Erlang VM is known for” or “everything Erlang is known for.”
I'm hoping to find a reason to use it soon.
I first reached for Elixir when Ruby couldn't handle large amounts of websocket messages. It really shines in high-concurrency contexts. I also love Phoenix LiveView and have a couple of side-projects running on it.
Do you have a program that doesn't need to run fast?
(wip, no guarantees, this is the engine i use)
https://github.com/ityonemo/ce_ce
For me as a long-term ruby user, though, elixir is not quite as elegant as it could or should have been. Even simple things such as "defmodule Xyz do" feels weird to me.
Love how dedicated you all are to providing clear entrypoints and being thoughtful around documentation!
Sorry, I meant the subjects (LLMs, ML, AI) are intertwined, not the people. But what I was getting at with my comment is that (IMO) most people see them as distinct things, even on HN where most know that LLMs use ML. As an analogy, it's like physics versus mathematics: separate subjects in most everyone's mind, and even separate academic departments, but physics is still math.
>Might as well use LLMs for the whole thing next time, since we will be accused of doing so anyway! :D
The only sloppy aspects that stand out to me are the needless animations/transitions.
But it's blazing fast which is good to see.
Hoping Elixir continues to thrive. It is such a great language (and such a great language for AI coding too!)
1. https://arxiv.org/pdf/2508.09101
OT: I wish more funding & development effort went into BEAM itself on making it more performant.
Note: I’m not talking concurrency. I’m talking pure raw performance.
Seems like it’s been a one person show for over a decade on making it faster.
I suspect once the Erlang/OTP team squeezes all performance in the JIT, they will look into optimizing across modules, which will probably open up many new possibilities, but it requires rethinking some runtime primitives.
Knowing what I know now, I would’ve tried to push for a threaded interpreter to get rid of the runtime overhead of dispatching altogether. I don’t know if they’ve changed the architecture of that module much since I left :-)
You’re an inspiration for many. Thank you.
I’m curious to know what your top 3 hopes for BEAM itself are for the coming years (in any area that you think would make it better).
1. The cross module optimizations I mentioned above 2. Have a WASM target for the runtime itself 3. Make it easier to ship single file executables with the whole VM
But they are really “nice-to-have”s. I have been a happy user for 15+ years!
You can get substantial performance improvements by using guards though. See what Wings3D does with is_float() everywhere in hot numeric-heavy code.
The URL: https://elixir.hexdocs.pm/getting-started.html
It's super handy. There's no security barrier between nodes. It's a headache if your network is unreliable.
For a chat app, messaging someone becomes a series of steps:
a) look up if they're online (send a message to the presence database service)
b) if you got a process id back, that's the process connected to the user, so send it the message. The process could be on the same machine or not, but the sending api is the same. This is the special part: few other environments make arbitrary messaging between processes/threads/tasks/whathaveyou so pervasive.
c) if you don't get a process id back, the user is offline; send the message to the offline database.
My usecase is less independent though, that control plane is orchestrating like Lambda/fly.io style workloads on top of firecracker: https://jomcgi.dev/ember
[1] https://www.youtube.com/watch?v=nZmDEUeHeVI
I was involved, years ago, in using Erlang on these devices: https://www.icare-world.com/us/product/icare-eidon/
It was a lot of fun and there were some very interesting challenges for everyone involved.
EDIT: shipped!
If you buy into the Elixir stack then you now have constraint you could've avoided entirely by avoiding it.
Also for devs there seems to be no premium offered for this talent pool scarcity. With LLMs I think language-specialists are redundant in a large scheme of things. ex) at one of my current remote jobs, I shipped an entire telecom infrastructure with barely knowing Elixir and we brought on contractors to audit the code and they found no issues.
> I shipped an entire telecom infrastructure with barely knowing Elixir and we brought on contractors to audit the code and they found no issues.
Erlang/Elixir experience is rare, because it's not widely used and the teams are small. It's not worth trying to hire for it. Hire for people who can figure it out on the go (amd are willing to give it a try).
You did it, hire other people who seem likely to be able to.
normal market dynamics suggest scarcity demand premiums but this is not the case with software developers it seems.
a) did you pay your Elixir contractors more than you would pay a Java contractor for similar work?
but also...
b) scarcity isn't the only factor in price. Erlang/Elixir developers are scarce, but Erlang/Elixir jobs are also scarce. You need both demand and scarcity to raise prices. Also, it doesn't cost much to turn a willing, good developer into an Erlang/Elixir developer; substitute goods reduce the impact of scarcity.
also c) if you found contractors, but not employees, maybe you weren't willing to pay enough... So maybe the price is higher than you thought?
Congrats on being one of the mythical developers that manages to get AI to write perfect code consistently!
if you treat it like any architecture then there's all sorts of techniques and knobs to produce consistent output
If i go full dynamic, why not use pure erlang instead?
I hope you don't use discord or rely on pagerduty.
I also worked on all of the copy myself, collecting feedback from core maintainers as I went. The new tagline was a suggestion from Theo which we iterated on. I did use LLMs as an assistant, but I did not ask it to generate the content.
Might as well use LLMs for the whole thing next time, since we will be accused of doing so anyway! :D
FWIW, I don't think the site fits the LLM template. The scroll through the use cases is particularly nice.
And thanks for Elixir. I love it, and the agent + tidewave loop is a joy to use!