Let’s say a company like Apple or Google gave you very large budget to go build something with Elixir - what would you build?
Can be:
If you want to share what you’d build with a large budget irrespective from where it came, feel free to say as well 
How large? Billions of dollars?
You didn’t mention that it had to make some profit, right?
I will happily create some Elixir apps to help to achieve world peace, to end starvation, to cure diseases, to solve climate issues, etc.
I’d probably spring for a collaborative modal terminal text editor! There are some real fun possibilities there with Elixir but the scope of the task is daunting as a side project alone.
I’d work on an extensive crossplatform functional-first graphics and GUI toolkit. It would work for everything from a low-level Nerves GUI to desktop applications (as an Electron replacement), and maybe even mobile apps.
I would build video processing tools, server side library, wrapper around ffmpeg.
Any reason why you’d write you GUI in Elixir, which seems like a very unusual choice? Why not write it in something else and then write elixir bindings for it?
Realistically I do agree with you but this is with a very large budget. And I’d prefer the world to have one Elixir GUI library rather than none.
How about a IoT / Serverless platform?
I think OTP/Elixir are a very good match for this use case considering, high availability, concurrency, scalability.
Non-Elixir related:
Elixir / Erlang:
I’d use Elixir/Phoenix to build a declarative tool that outputs HTML. As an end user, I don’t want to “design” an HTML page (design, to me, includes HTML node hierarchy as well as CSS and layout). I want to select, multiple-choice style, what information I have and what I want the page to accomplish. I want to fill in that information, and then be prompted by several finished outputs that I can choose between (I think humans are better at reacting to what we like/dislike than making decisions from scratch). Then I want to click “publish” and get a live URL. No servers, no creative design decisions. I think this would be far superior to the complicated mess that is drag-and-drop HTML builders.
With an unlimited budget I would create a business analysis tool that is specifically targeted at Small to Medium sized businesses to act as a digital CFO. A number of the available tools require a fair amount of knowledge to maintain and operate and I want to make it super simple to use to actionable insights.
An OS, with Elixir all the way down (TCP stack in Elixir, anyone?) and up (I already started on the GUI bits ;-)).
A Nerves-powered moon lander.
mic drop
A universal mobile network. I would like to speak with more aliens about their budget, how they would like to spend money, how they think they could get others to pay for their ideas and what/who blocks them to escape their prisons and get some bright ideas.
An event-sourced CRM with declarative/point-and-click configuration such as Salesforce. If I had a moderately large budget ($30-200 million). I’ve done a lot of work with Salesforce, and there’s a more than a few good reasons for why I’m spending an increasing amount of time with Elixir…
Basically I’d like to have the capabilities of an event-sourced system built with highly composable components. I’m confident this composability of components/contexts/domains is only possible with event-based integration and functional language features. Declarative controls to manage workflows without the need for a dependency on DDD/CQRS/Elixir knowledge/skills is just the requirement to get users for a tool like this.
I would work on Blockchain-like global immutable pseudo-identity global database where people can maintain N virtual personas with zero personally identifiable information attached. See above: if this is combined with a new stateless protocol that doesn’t even track your IP then this will IMO save the internet from the merciless profiling and personal data harvesting that it suffers from today. (Well, it can open the doors for a huge wave of trolls as well but that’s happening right now already.)
One of the reasons I want these sorts of business information systems to be event-sourced is so that they’re natively compatible with blockchain paradigms. I’d like the “user” to be a separate, distributed, decentralized, and secure identity that is only authorized by an organization/system. There should be one identity/user that an application/smart-contract can utilize instead of every central service having their own user, login, profiles, etc. (Although there should probably also be pseudonym features for obvious reasons.)
Say the information systems people work with are all event-sourced, and every person has some kind of aforementioned user-aggregate. When you participate with an external system it stores it’s own copy of the system event you’re participating in, but since you are part of that system event, it is also appended to your user-aggregate.
At this point we could implement some interesting authentication mechanisms by transmitting encrypted copies of your user-aggregate events to one party that they can then verify with the other third-party’s system that produced that event. This prevents the need for every information system you participate in from viewing potentially private events on your user-aggregate.
Finally this user-aggregate could be used as a signal of merit (e.g. Resume/LinkedIn page). A service that has a sufficient data-set of merit-relevant data could make estimates as to your capabilities/qualifications/subject-matter-expertise. Again this functionality could be implemented without exposing your private details. When there’s a sufficient trust of these merit-summary services, we add a lot of capabilities to our economy such as merit-weighted-voting, gig-economy improvements due to reduced hiring costs, and actionable data to tackle fake-news. In my opinion the large scale systemic problems we have in society are rooted in this problem we have in estimating the credibility and capability or people and organizations.
So I guess, with an unlimited budget, I’d work on solving the problems of merit, identity, and governance because they’re all connected and would have the biggest impact. I suspect this sort of problem can only be tackled with an unlimited budget. 
I would build a distributed shell environment, such that machines could be serviced and upgraded without me having to restart my ssh session, re-establish my screen session, start editors and other context. It would be managed like a highly available platform. It would minimally include a distributed shell environment, a distributed editor.
And since the budget is unlimited, I would also build a data system (database sort of thing) that is a purely functional environment so requests to the system would be passed through a set of pipelines similar to phoenix, which would be “plugable” in such a way that a very simple or complex distributed data system could be built in a modular way.
I think a small even $10/month donation from 2K people would already enable the creator to work on the thing full time, so I think even without Apple’s $ we can have an impact.
I’d stop wasting $ on Go and invest equivalent resources in Erlang/Elixir 
A brain. Each neuron equals one process. Talk to neural specialists on how to connect the processes, put endless computing power in it, connect sensors and start to teach it everything. 
QUANTUM BEAM VM 
© Copyright Elixir Forum | Terms | Privacy & Cookies
