GigCityElixir 25 | Dan Janowski: BEAM into the Future

[Music] Good afternoon. Uh, thank you for the introduction. I'm Dan Janowski. Um, so, uh, let's see. I could skip ahead. Uh, the one thing that was left out of my bio is, uh, I occasionally race a 1992 Miata. Um let's see. Okay. So by the end of this talk I hope to have you convinced that we need to act as a community and that each of you will get involved. So um what's really interesting about uh we seem to have a theme today in terms of computing and and the subject of talk. So that there actually is an interesting amount of overlap in what I'll I'll say but I'm going to a different destination. All right. So first I'd like to say a few words about the computing where we are in the computing timeline and the implications of our evolving AI present. So in less than a 100red years the once esoteric computer has transformed the workforce and the global economy. It has accelerated automation and manufacturing productivity and is now automating technology itself. The technology that none of its creators can explain and is trained on data stolen from the internet and of unknown quality. AI is here and while it gets better, capitalists will try to undermine the value of developers and try to make us replaceable labor. Our stream of junior engineers may shrink as the jobs won't be there anymore and the pressure for programming productivity may become as dire as a fulfillment warehouse. Of course, this is not necessarily the future, not a foregone conclusion, but the makings of it are already here. Uh so there are risks where we are at this moment as it pertains to AI. There's a lot of capital allocated to AI. If we are indeed on the incline of the hype cycle, heading towards the peak of inflated expectations, then there's going to be some pretty terrible financial fallout from a crash. We know that nations are in an arms race for AI, and that will probably get hotter and lead to more geopolitical tensions with side effects that are hard to predict. In the near term, there are projections that AI will need up to three times more electricity than we currently produce. This will increase all energy costs. It will also increase climate change pressure as a lot of that generation will be with natural gas. It will also divert resources away from the energy transition and humanity's efforts to stabilize climate change consequences. So these are challenging times we live in. Occupations that we've enjoyed for decades now may be under threat and the prospects of what AI will do to software development are hard to calculate and potentially disastrous. We may forget how to do things and end up relying on AI for all of software development. Either way, the effect of AI domination will be a larger further shift in economic disparity and the concentration of both information and power into still fewer hands. But I believe the world needs Elixir. The language is productive, powerful, efficient, and a pleasure to use. It's a way to run software that matters. It's the OTP way. There are less things to deploy and that makes things easier. Elixir is also a great integration platform and Explorer is a great example of that. Elixir is has a very approachable functional programming model and that matters because functional programming makes better software more reliable, more understandable and safer. And we want Elixir. We want to use Elixir. I know I want to use Elixir. It's fun to use and I always learn from it. I've become a better programmer because of Elixir. I eagerly look forward to using it. I know that it's a tool that I can depend on. It doesn't do unpredictable things. It doesn't give me a headache. I can read it with a natural ease that no other language compares to. I have never found myself at a dead end. There are always approaches to solving problems. I've used many languages over my career and Elixir longer than most and I would not choose to use something else at this point and macros of course. But we need to help Elixir grow. Organic growth has its limits and we do not have a mega corp advocating for it. It's not a typical school language. It's not shipped with the operating system. When you look at vendor API support, Elixir is not an example or a supported client. And functional programming is also not automatic for new people. And we can help Elixir grow. We need to work on building awareness and acceptance with decision makers like founders, CTOs and product owners. We need to exchange with uh engage with developers outside our ecosystem to plant the seeds of possibility. We also need resources to help with that advoc advocacy to speak clearly and with authority because it's not easy talking to others in convincing ways in their own language. AI can act as a bridge and an accelerator. So, if you have not read Beating the Averages, which is an article that Paul Graham wrote in about 2001, it is still relevant today because it describes how a small team with good tools can compete and succeed. Um, how many people have read this article or know what I'm talking about? Okay, a few. All right, so I'm going to summarize. So Paul Graham and Robert Morris started a company called Via Webb in the late 90s. It was a novel online store builder al um that ultimately became Yahoo stores. When the competitors showed up, they used techniques that were common at the time, but via web was written in lisp and gave Graham and Morris the agility to innovate faster than any competitor. Their technology choice made the difference. And while Elixir is not lisp, it has those kinds of advantages. A lot of technology choices are not made based on the quality of the technology, but for others other sometimes ridiculous reasons. Paul Graham showed that making a good technology choice changes the game and elixir does enable agility in the right hands and it makes a strong business case. So we need to work on acceptance with decision makers with a message in clear business terms that they can understand. We need to appear in different places because it's harder to access seuite people and minds. We need to reach them before they get there playing the long game. Founders are easier because they are looking for an edge. However, middle management operates on inverse incentives where they're looking for bigger budgets and headcounts. And for that reason, a technology that reduces both is a harder cell. I suggest that the opposite is true in the boardroom where profitability does matter. If profitability is possible with a small team using Elixir, it can work for a large company just the same way. So if you listen to thinking elixir the podcast any hands? Yeah. All right. Excellent. Uh so you may have heard this an interview with Michael Lubis u from Paraxial. Um and this is specifically about security but it summarizes sort of where we are as developers. We are so happy writing Elixir stuff that we don't really think about what the CTO needs to think about or what their concerns are. And they're concerned about security. They're concerned about industry standards. They're ex concerned about liability. So if you haven't listened to it, I do recommend. Um but he because he talks generally about the importance of security, but he makes this point and I think this is worth keeping in mind. It's really um appealing. It's really comfortable to be working in the space that we love. But if we stay there and don't think about how the rest of the world sees us, we do it at our peril. So you may recognize this table, Sasha. >> You've seen it before. Yeah. So what's really funny about this table is that it hasn't actually changed that much. I mean, server A is maybe cloud services now when, you know, when it wasn't, but the complexity of of other technologies in terms of the galaxy of other pieces that you need in order to make something that's reliable is complicated on its own. And we can still do all of those things. We can do more than when he originally wrote this. Um, there are some really good examples of what of how far you can take this and water park is one of them because no database, none of this other stuff. It's all the beam. It's pretty awesome. If you have not watched that uh or didn't see that presentation, I think it was last year at two years ago here. Totally worth it. It's like a It's like a fever dream about like how you can do like amazing things with the beam. Oop, not that. Okay. So, here is a simplified diagram of uh like a problem assessment and where we need to think about inserting ourselves or doing a better job. Okay. So when people have a problem and they go have to go searching for an answer because they've run out of road with whatever technology they're using um they'll go searching for answers and if they can find elixir or they've heard of elixir before then we need to be there with resources that guide them through an assessment. Can Elixir solve their problems? We have lots of case studies and articles published already but I think we can enhance this if we direct directly address the kinds of problems that crop up in other ecosystems. So think very specifically like Ruby has this kind of problem or Python has this kind of problem and it's a common sort of scaling problem whatever that is and how we can address how we handle those problems. So if we formulate solutions uh to those kind of problems that can be quickly understood then people may give us a deeper look. And I'm talking about people that don't know anything about Elixir. Maybe they just seen it or heard of it. So the next critical step is for someone to assess whether it's going to be hard for them to learn and how how much it will take to build a solution. So if we have again those specific kinds of examples from one language into Elixir, we can illustrate a way for them to make that transition. And we also want to address some of the aspects of functional lang functional programming because that will trip people up. If we can get them out of this face with some optimism, then the next question we have to answer for them because anyone who's making this kind of assessment will ask, how is this going to fail later? I we've all probably done technology assessments and we look at the kind of glossy first page and be like, wow, that's really awesome, but how's this going to screw me later? So I think we have very concrete things to say about this and um because Elixir solves a lot of problems on its own but that we're never go people are never going to be stuck. There's always options. So if we can get them through this phase then there's lots of books uh that people can read. Um there are lots of courses. I hear someone here maybe gives courses. Graio. Um, so with each of these stages, we need materials that are very specific to the purpose, helping people that are both technical and non-technical make this kind of assessment because we don't know who's going to go looking. Okay. So, resources, um, we need to enable telling the story, uh, the elixir story by a larger group of people within the community. And what I what I mean is we have to enable the community to tell this story better. So by developing resources that helps people in that process, we need to have guidance on how to frame advocacy talks and articles and posts because we can't use the same mindset that we use when we talk to each other. um comparison sheets um and sheets that sort of like debunk misconceptions um using sort of a quantifying uh approach to success stories uh and in different forms like from an elevator pitch to an in-depth analysis. Uh I also think that we need a marketing site that is specifically focused to guide travelers into the elixir space. So AI can be a bridge uh if we use it carefully. It is just a tool ultimately but uh but but it can be a bridge for this purpose. So currently the Erlang ecosystem foundation education working group is working on a set of fine t fine-tuned domain specific tasks in elixir Erlang and functional programming. Um, X Moralda is something I like tripped over the other day, which is a rag that basically points at Hex. You can ask about any package on Hex and get like actually quite usable answers. Um, Tidewave was just released using MCP. Uh, the great thing about AI is that we can take someone's quest or parts of it and translate it into Elixir code. So we can answer people's questions like very specifically so people can get a sense of how they might solve the problem and limits etc. Um but we can put people on the right path and it can also help with again with the same functional programming transition. If we fine-tune things right it will encourage also the right idiomatic style. And this is what we all come to is what can you do because we all need to do this together. We all have we all work in elixir or we have interest in elixir. We work on products. We all have something to contribute and a lot of us are very happy to my earlier comment just doing the thing that we love. But we have to think about how we can help Elixir grow because the world is changing. both on the AI front but also in just the general climate and we want to do everything we can to ensure that the thing that we enjoy doing we can keep doing. So, here are some specific things you can do. If you are not a member of the Erlang Ecosystem Foundation, please join. It is free. If you can afford the $100 a month uh me or $100 a year, sorry, membership, please do that. Um there are working groups for all sorts of things. I am actually a member of the education working group and the security working group. There are exceptional people that are trying to make a difference and your help can make a difference. If your company can be a sponsor of the Erlang ecosystem foundation or participate in grants, they fund the EF funds stipens to get some work done by dedicated people. So for instance, we have a person who is working based on a stipend that is dealing with a lot of security related issues that elixir and the whole ecosystem need to have handled or we are going to be cut out of projects because of regulatory requirements or corporate governance requirements. So we have uh we're making great progress with software bill of materials. We are also setting up or have almost completely set up our own cert numbering authority. So we have a pipeline for dealing with exploit um reporting. Um there's the uh yeah we have the CRA which is coming in Europe. There will be similar things globally. So if we don't do this we run the risk of being excluded. And this should be easy, but we don't do it enough, I think. And that is to be active in public, posting, writing articles, whether it's success or failure. There is a survey. Have any of you participated in the Elixir survey? I haven't. I I only just learned about it. So, it's I think it's only maybe been posted on Elixir forum, but this year I'm going to make a point that it gets marketed like much more broadly. I talked to the guy that runs it at Curiosum and they are going to run it again this year. They only got 500 respondents last year, but the more people that respond, the better the data, but also it makes a much better statement about the people engaged in the community. And that's really what this is about is being visible because when people come looking at Elixir and trying to evaluate like is this a vibrant community or not, these are going to be the first things that they see. They're going to look up Elixir in on all these different places. Okay. So, we have lots of resources to create based on what I've already said. We need cross language solution papers. um pro if you can profile your company or your product, that's a really valuable thing to do. Like there's a whole list of uh companies that use Elixir or use or or in the ecosystem, but they don't go far enough in talking about how they use it. There are some amazing stories that I've heard that are not publicized that are used in like pretty big, you know, uh, deployments. Um, so if you can do that, if you can do it in written form, either by a blog post or or some other kind of article, that's great. But there's also a bunch of podcasts and it makes really interesting material to hear people talking about how they're using Elixir in production um or how they evolved their use of Elixir in a podcast because it can be both inspiring but also for from an adoption standpoint if people come looking they will hear the success stories and the enthusiasm for what we do. Um, so I made mention of a website. Everyone's got a website, but I think we need a website that is, if you look at the Elixir Lang website, the EF website, they're great. There there's lots of links. There's lots of resources. It's all very navigable, but it's really meant for the community, for the people that are already here. But we need something that will guide people who are just showing up. Like in the strict sense, if we are successful in getting out into the world and saying, "Oh, I used Elixir for doing this." Where are those people going to go to say, "Oh, I just saw Elixir." Like, what is it? Like, how is it going to solve my problem? Like, why should I care? And that's the thing that I think is missing. I haven't found yet. And maybe it's out there, but it's not obvious. it doesn't come up in in any of the places that I've looked. Uh let's see. Okay, so there's lots of pitfalls in making this kind of content. It's easy to get it wrong. So the more the community is involved, the more people are looking at the content, the not only the more of it there will there be, but the more accurate and more complete. So there's a lot of value to more people being involved. And for those of you that are courageous enough, get out there, go to conferences that are outside the Beam ecosystem and talk about what you do. Now, this is not easy. I don't say this like flippantly because talking about Elixir to people that don't know anything about it is not straightforward. We can't use the same talks that we use here to talk to people about why they should care about Elixir. But we can go out and advocate at meetings and conferences. And we need to be able to illustrate how Elixir solves problems and provides better answers. We need to challenge people's assumptions and allay their reservations, but we also want to excite people. And we need these in both technical and non-technical talks. And the non-technical talks will help business people understand the value proposition of Elixir. This is a great conversation. Uh this was actually recorded at a Ruby conference in 2024 and the inter it's a sort of a interview style thing but Jose talks about the bootstrapping of elixir which has been talked about already today but that he went out I mean think about it he invents elixir and then how is he socializing it like he's literally got to go to conferences like Ruby conferences and talk about a language that nobody's ever heard of and convince them that there's some value in them rethinking like the way they do stuff. Now, he has lots of advantages that we don't have. He's obviously the creator creator of Elixir and he was he was part of the core team for Rails. So, he had notoriety in the Ruby world. Um, so he's actually invented to some of these things. But even when he's spoken, the audiences are not really big. But if we can reach a few people and change a few minds or create the kind of curiosity, the community can grow. So depending on your industry or your sense of adventure, you may be able to do this. If you can see a way to speak about elixir in non- elixir content contexts, meetups or conferences, please go and do that. Now I started this paper or this this talk like a month ago and in the course of that month so many things have changed. I've had to change this talk so many times between AI and now this this is coming out next week. It's uh it's it's a sponsored program by Dashbit and Oon Pro that if you want to go and talk at a conference outside the Beam ecosystem within your region, there is there is money to pay for some of your expenses. Uh I don't know exactly when it's going to be announced or how it's going to be announced in print, but watch for it if you're interested in doing this. So we're all working on products and at companies that are not just doing Elixir for Elixir. We're all crossing into into actual industries. So the these are just some of the venues to consider. But if you're working on IoT projects, there are places you can go that are not within the Beams ecosystem to talk about what you're doing and creating that moment of, oh, what's that? And it doesn't even need to be like uh all about elixir as like a central topic, but as long as you're making the the statement out there that we're using elixir successfully and like these are the upsides, it will create some interest. Okay. So in 1735, Benjamin Franklin wrote in the Pennsylvania Gazette to warn Philadelphiaians whom he felt did not take fire prevention seriously enough that moving hot coals around the house without care, dropping floor embers into floor cracks might result in having to jump out of your window after midnight and hazard breaking your neck. Um his point was to anticipate or risk the consequences. So a wonderful feature of anticipation is the time it affords. There's no panic. So coolheaded planning is possible. Once an event has come to pass, there's never enough time. And as a community, we are planners. It is part of our work. So I'm going to paraphrase Richard Clark uh who was the director of counterterrorism in 2000, a very pivotal year. So the problem with preventing disasters is that no one ever knows because nothing ever happened. And this feeds a false sense of security and doubt of dire warnings. And we as human beings have a tendency to warn to ignore warning signs or as a derivative think someone else will deal with it. Jose has nurtured elixir and advocated broadly and we benefit from Ericson's commitment. the many commercial and volunteer contributions and conference organizers. Elixir does not have a mega corp looking after its ecosystem. But that also means we don't have to submit to the whims of one either. This means growing and protecting elixir is our collective responsibility. If we are collectively, if we all collectively advocate by building advocacy materials, by contributing to an active society online and in real life, and by representing Elixir to the larger world, we can ensure it will continue to grow and be the vibrant community we all want to be a part of. It's up to up to us to build the future we want. And I believe if we act upon the things I've talked about, we can make that future happen. We can do it together and be better and stronger as a community. And I hope you see the possibility and opportunity and will join the pursuit. Thank you for your time.

gigcityelixir.com Beginner, hobbyist, or if you use Elixir in your day job, you will find talks here that change the way you look at code. GigCityElixir is brought to you by Grox.io #ElixirLang #GigCityElixir