You’re right about the negative effects stemming from lack of big tech adoption, but I worry about the negative effects if this ecosystem gets the said big tech adoption. It would likely be a classic Faustian deal.
dimitarvp
sodapopcan
Was about to reply the same as @dimitarvp. If Elixir gained a massive audience, there would likely be a lot of pressure to change the language (and ecosystem) in drastic ways. I’m not going to give any specific technical examples but you if you read this form often enough you’d see the fairly extreme hostility and vitriol of people trying out the language who feel that it should accommodate their idea of how they think it should work. I’m personally fine with leaving these people behind. At best we get called gatekeepers, at worst we get called cultists. And in true cult member form I Trust The Core Team 
and it’s ultimately think it’s up to what José wants.
2
hubertlepicki
What if these two things are related and the “extreme hostility” is actually a valuable feedback that needs to be incorporated or at least considered in order for a language to have a broad success? A general purpose language needs to be able to solve multitude of tasks, and be easy enough to get on so that developers can be trained to use it relatively easy and cheaply. Again, it’s easier to interact with when that feedback comes from a corporate stakeholder in a suit rather than an autistic random writing from his mom’s basement. I know for a fact that the later can be and is frustrating to maintainers of open source software.
0
dimitarvp
I almost would believe that if I didn’t see plenty of toxic examples of people straight up trashing the syntax or bemoaning the lack of their favorite pet OOP feature. And loudly proclaiming Elixir is “trash” for not having it. I can and have taken loads of feedback that I used to better myself in this life but those kinds of people don’t seem interesting or useful in this regard.
For one reason or another – and I am very surprised by this – even having a very casual-friendly FP features like Elixir does (f.ex. immutability) gets people severely confused, almost helpless even. I am not sure what we can do except pack shop and say “okay boys, we are a Ruby now!”. Do you see any other way?
Eh, if you say so. They are difficult in different ways is what I’ve observed. The corporate stakeholder can listen to you for full 45 minutes, then bite off his cigar and then calmly say “Very nice, now please implement this or there’s no next funding series”. 
If you are implying that they are more reasonable… I have my doubts.
Still, what would you change in Elixir?
My absolute top #1 with zero competition would be to start shoving static typing in Elixir as quicker as we can. In all my Elixir work, having to add more and more tests and carefully reason about return values and making sure we’re not operating on a string when we need an integer or a float, has been the chief problem.
But I am known to be unrealistically idealistic at times. So let’s dial this down.
I’d probably work hard – if I was paid to – on making Elixir less “raw”. There are many things in the ecosystem that are made perfectly to accommodate other libraries… but not always apps. Ecto and LiveView come to mind where one abstraction level up would make Elixir feel much less manual to be perfect for just quickly making apps.
I appreciate how Elixir carefully focuses on providing great building blocks. But… nobody is willing to put the work to make working parts of apps that can be reused, out of these building blocks (myself included). Everyone discovers and rediscovers the same lessons as they are building their app.
To me that feels wasteful and I would work to change it… though thinking further of it, why would I give shortcuts to greedy lazy people is another entire dimension in this multi-dimensional matrix. Feels like we’d go too far into the “automate myself out of the job” thing, too far for my taste. Maybe.
But yeah. Static typing. If Rust was not so verbose and wasn’t still figuring out ORMs and DataMappers and web frameworks, I’d probably be a long time regular Rust dev by now. Elixir has mature choices that, even with some flaws, are pretty much perfect at what they do.
2
hubertlepicki
dimitarvp
I had polite people beat me up until my ribs almost cracked, some 25+ years ago, just saying.
0
sodapopcan
You put that in quotes but I was not exaggerating. The accounts I’m talking about have mostly been banned or locked to TL1 status with public warnings from mods.
I don’t view Elixir a “true” general purpose language and I think that is one of its strengths.
Again it really hinges on the goals of José and co, but I would bet on Elixir surviving a long time as the secret weapon of anyone willing to invest in it.
Ash 
5
dimitarvp
Guilty as charged. 
I just can’t seem to ever get to it… Definitely a bigger and bigger blind spot in my field of view.
0
regex.sh
sodapopcan
That it trades fast IO for slow CPU to the point that NIFs are a requirement. It’s not hard to get into a situation where you need one or three (I haven’t worked on a production app that doesn’t have them). I sorta see BEAM languages as DSLs for creating client/server applications that also includes typical language features and a massive standard library. That may not be a totally accurate but I’ve heard other people express roughly the same sentiment 
2