Want to get all the different ways folks are pitching Nix and see if we can collect all of it into the ultimate “How to pitch/talk about Nix to the non-nix aware”.
So, I had an interesting but probably not super unique experience at DevOps Days Chicago. I was sharing the magic of Nix when someone, new to Nix but deeply fond of it, expressed a challenge: how to effectively convey its beauty to others who haven’t dived in yet. It got me thinking - can we brain meld the ultimate “How to pitch/talk about Nix to the non-nix aware” guide? and maybe share it at NixCon!
Brainstorm Questions -
What do you say if you had 30 seconds?
And if you had a bit more time (say 120 seconds)?
They jump to the Docker question.
What’s going to change in my life tomorrow if I start using Nix?
Not like I manage to convert that many people to Nix/NixOS but here’s my two cents.
deeply fond of it
A first advice I would have would be to take a step back. Too often I see passionate people fail to “sell” their passion. Pitching is not about yourself but about others. My belief is that the conversation can’t start with “I love Nix”. It has to start from the other party. “What are your struggles?”. “Did X ever happen to you?”.
How to pitch/talk about Nix to the non-nix aware?
It is known that the first few seconds of a pitch are crucial. Something fun to do is to watch a show like “Dragon’s Den”, listen to the first few seconds of pitches, list the ones that best catch your interest, and find the techniques that they use. Or ask ChatGPT: “Give me a short list of 10 advice on how I should start a pitch”
As I’ve said in my NixCon talk last year, I think that telling a story is the most powerful technique.
It becomes really effective when that story connects with the people in the room.
And that depends on the audience, but also on yourself and your own experience.
So when it’s a 1:1 pitch I ask the other person questions to find a struggle they have that Nix/NixOS could solve. If it’s to a crowd it’s harder, but usually the crowd has a common background, so I would do the same: prior to the presentation I would find a recurring theme about something they care about that Nix can solve.
What do you say if you had 30 seconds?
what is the audience? the context? your goal?
They jump to the Docker question.
If they have “enough” experience with Docker they know that building the same image months apart can fail. This helps introduce reproducibility. If they still don’t care it’s no use.
What’s going to change in my life tomorrow if I start using Nix?
It pretty heavily depends on who you’re talking to. For instance I had a mathematician friend who uses Arch, and my pitch to him was that Nix treats packages like a directed acyclic graph where each node is hashed based on its contents and the hashes of the nodes it points to. We talked for a little while about the graph theory and cryptography of that, and by the end of the conversation he was extremely interested.
That pitch would not have worked with someone who wasn’t a mathematician.
Nix / NixOS solves way too many problems to pitch it to everybody, as you are more likely to pitch the wrong way which might let people think that this is too good to be true. Doing pitches to random people is the best way to bring discredit on whatever you say in the future.
Instead of pitching, you should be listening to what problems people are facing and then ponder whether using Nix, for this one use case, would solve their problem. Then pitch Nix only for this one use case.
At least that’s my experience from the past decade, and hopefully this is changing as more and more people are aware of it now.
When I have to pitch NixOS it is usually not “why Nix”, but “why not Debian”. My answer is that Debain or any other mainstream Linux distro requires years of experience to use properly and the processes that an administrator has to go through involves a lot of unwritten “this is just how its done” details. I point out that I don’t see why a NixOS admin needs more specific knowledge than a Debian admin, and that by sticking to the distro status quo you also commit yourself to perpetuating the status quo of who gets to be a Linux administrator. The burden of proof that our distro is better than the others is certainly on us, but I think we settled that years ago.
I totally agree with what’s been said here, especially the main idea of first figuring out the challenges or “pain points” our listener is facing. Then we can see how Nix might tackle one or more of those problems. If Nix can solve multiple issues, it makes sense to focus on the one that seems most important, just like @nbp, @Arsleust and others pointed out.
Doing all of this on the spot can be pretty tough! It’s all about understanding what the listener needs and explaining how Nix can help.
I’d love to hear more examples from you all. If you’ve talked to different people and found out what their unique challenges were, please share what clicked with them. Stories like “When I was talking to (type of persona), they were struggling with (challenge), and what really connected was (solution)” would be super helpful.
Side note, also loved the example from @hansemschnokeloch on how simple it can be.
I wouldn’t pitch Nixos as is.
I think they need to chose flakes, and then take all the documentation from all the disparate sources, give ChatGPT a few parameters, and then tell it to write a Handbook.
Every person will care about Nix for a different reason.
So I try to write Nix code to solve people’s problems. Then they do a single Nix command, and it does what they have previously spent hours of manual and error prone work doing before Nix. Doing this takes me 30+ hours, unpaid, every single week. This is an investment.
To impress people, all you have to do is solve their problems in a Nix way, give them a Flake, and have them realise that it can automate 99.9% of their problem away, on their own turf. They know the problems in their own space better than anyone, and they will appreciate the solution you provide and hopefully be interested in how it was accomplished.
If they don’t accept it at that point, they will accept it in months as they realise it is more difficult to do things without Nix.
I’d you use Linux and want to convince someone to use it instead of Windows, all you have to do is use it whilst they’re watching you solve the same problems, and they’ll notice that you don’t have as much pain. They will decide to use it on their own if they want to. I think the same applies to Nix.
TL;DR if you write the code, make it available, demonstrate how it solves a problem, it is the best pitch you can have.
Here’s an example of a Nix function I made to help the sanity of a friend who is employed. It took me about 6 months to figure it out, and it will help convince anyone who is trying to build embedded products with Zephyr RTOS. It eliminates the need for tons of incorrect, misleading or obsolete documentation upstream and simply solves a problem they have of managing their toolchain. I’ll shortly make a video about it and demonstrate how to use it, and we’ll see how effective that is as a pitch.
Remember SVN? Nix is the Git of dependency management.
It seemed true in the way of thinking, the steep learning curve and the less than ideal UX .
I also tend to compare Nix to lock files. Lock files are known and seen as obviously a good thing. Usually there is a lock file ‘with all dependencies’ of a project, but the readme still asks to install more. Shouldn’t those be locked too? Nix can!
In that same vein I’m going to give a presentation at work that is titled Nix: locking dependencies all the way down.
Not really a pitch for Nix itself, but I found that when I got asked about “why are you using XYZ” and that I replied “Don’t bother, you won’t like it”, people get curious and often end up using it. And what’s cool with this method is that I didn’t promise anything so I don’t become responsible for any question about that XYZ or for any failure of using it.
This may not be the best way to market that though
But I have to admit, if I saw a message like “The package manager that isn’t for you” on nixos.org, I’d be really curious
Would be great to have a practical analysis of things in production which traditional distributions are doing bad and NixOS is doing great or vice versa, e.g. updating software that isn’t in the official repositories or updating Let’s Encrypt certificates. I’m using NixOS for five years now and don’t remember all that little things anymore
but here’s my two cents.
