People want products, not projects

There's a large feeling in my industry that the internet is getting more and more monotonous. Hand-crafted works of internet whimsy have faded towards the same drab design that everyone uses. It becomes merely populated with people's information rather than people truly customizing it. As things get more complicated and scale to more users, options for individuality and creativity get snuffed away in the names of consistency, security, and so that the people making those platforms don't have to test as many invariants.

This is understandable, but it sucks, but it's understandable.

There's been a call towards moving people away from products like Twitter in favor of projects like Mastodon, but I feel that these fundamentally miss the point of why people use products in the first place. This fundamental misunderstanding leads some projects to be seen in the same vein as products, but ultimately make things fail to land with people.

Products and Projects

These two words are things that feel like they should be identical, but they are acutally very fundamentally different. I'm also a software person by trade, so please understand everything here as talking about the software industry.

A project is something that is created by a group of people that aims to solve a goal they face and then stops there. Typically the results of projects are for solving individual goals or the needs of smaller groups of people.

A good example of a project is Mastodon. Mastodon is a project that gives you most of what people like about using Twitter without having to actually use Twitter. You can post things, repost things other people said, reply to things people have posted, and view feeds of other random things other people have posted. This gets you most of the same experience as using Twitter, but without the downsides of actually using Twitter (lax moderation due to the sheer scale, etc.).

In contrast, a product is something that you can get "off the shelf". A lot of the options have been pared down or sanded away so that things just work. There's a core philosophy at work there powered by a team that spends a lot of time figuring out what the best way for them to do the thing is.

This is different from a project because a product can be adopted without many additional steps like setting up infrastructure, purchasing a domain name, or knowing what an Ubuntu is. Think about the number of companies that managed to get internet presence when the complexity was reduced from "pay a bunch of money to find this weird nerd to set this up for us" to "open facebook.com, make a page, share the page". You don't have to know what HTML, TLS, Apache, or anything is to make a page on Facebook.

One way you can conceptualize this is that a project needs work to fit into your needs, while a product can easily be adopted without any work at all being required. Running a Mastodon server for a community is a project, but Twitter as a product is there and you can create an account in less than an hour. Unless you really know what you are doing, it's inconceiveable that some random person could really figure out all the technology and maintenance schedules needed to set up a Mastodon server.

Similarly, choosing a homeserver to use is a problem that Mastodon has that Twitter never will. This is core to the Mastodon project, but never will be relevant to the Twitter product.

Fun fact: this is the same basic friction as the "protocols vs platforms" schtick that people were on about a while ago. Protocols require work to adapt them to existing molds. Platforms and products that let you build things on top of them. Compare IRC and Discord. IRC is a protocol that's been around about as long as I've been alive, and it's historically been used by open source projects all across the world to coordinate projects that quite literally keep a large amount of society together. However the protocol has stagnated and every attempt to modernize it fails because some random person doesn't want things to change. Meanwhile Discord as a platform is able to onboard more users per month than many IRC networks ever had at their peak user count.

I can't reply to a Linux Kernel Mailing List thread from my iPhone but I sure as hell can comment on a GitHub thread from the same device!

Why can't they just take up a project?

Some of you in the audience are probably thinking "wow but why can't those people just learn how to set up an email server instead of paying Google? It's not a problem for me. I have many rare funko pops." People use products because projects require work. They require you to have a vision. They need you do consider a design. They take actual real work in ways that I expect that many people don't have available because life is fucking excruciating. The last few years have burnt people out to a crisp. Being able to enter in the name of the business, the address, some photos of the restaurant and open hours is so much easier than having to have any of the opinions needed to set up a website that it's not even funny.

Communities don't want to maintain infrastructure, they want to talk about their speedrunning misadventures.

People don't want to have to know what an instance or a homeserver is, they just want to talk to people.

Nobody really wants to set up a bouncer to stay in their chatrooms for them, they just want to scroll up and see what other people said.

And finally, why should you have to set up a server, learn what Docker is, understand what a postgres is, and all of the errata with DNS, Let's Encrypt, or uptime is when you can outsource that to Twitter.

I guess the real consequence of relying on these commercial platforms is that the floor can and will fall out from under you in a moment's notice. Skype went to shit without much warning. Twitter did too. Discord will next. It's not a matter of if, but when with these products.

Edge cases

Of course, because I set rules in this article, there are exceptions. These exceptions can get kinda hairy, and both of them really push the lines between product and project.

Bluesky

Bluesky is the main company behind the AT Protocol, a protocol designed to give people the same experience of using Twitter while allowing for people to migrate their data and identity between homeservers at will. The exact federation UX hasn't totally been nailed down yet, but the basic model is that people have domain names as their usernames. I'm @friendshipcastle.zip thanks to this JSON route (powered by this code). You can use DNS though.

The real weird part about Bluesky is that unlike Mastodon and the fediverse at large is that their federation support is not implemented and deployed yet. They also have some questionable views on how moderation should work which is very obviously well-intentioned but definitely not from the perspective of someone who has suffered at the hands of organized harassment.

People are using it and treating it like "the new Twitter". There is also a culture there that seems to be treating it like it is a centralized service and not understanding that everything they post there is public. Very public. Extremely public. It's easy to understand why people think that it's private though. Here's a link to one of my skeets (yes, people really do use slang for ejaculation as the term for posting to Bluesky): staging.bsky.app/profile/friendshipcastle.zip/post/3jvu2krd6am2j. If you have a Bluesky account, you'll be able to see it. If you don't, you can't. This confusion happens because their main app is invite-only, but the undelying protocol makes things public by design. You can change the staging.bsky.app domain to psky.app and then you can instantly see every skeet (god I hate that term, I'm going to use "post" for the rest of this article) with curl:

<meta name="twitter:card" content="summary_large_image" />
<meta name="twitter:title" content="Twi (@friendshipcastle.zip) " />
<meta name="twitter:image" content="None" />
<meta name="twitter:creator" content="@Twi" />
<meta property="og:description" content="What's up gamers?" />

They're public even if the website doesn't show them as public. I can't blame users for treating it like a platform and assuming people can't read their posts without an account because that's how every invite-only platform works. I guess it's a protocol and a product, but that's a weird class of thing that we haven't really seen before this.

God knows what the fallout is going to be when Bluesky enables federation and right-wing assholes make their own homeserver and use that to bypass earned bans. All of the Bluesky's mantra of "you don't need to block homeservers" doesn't bode well from an ActivityPub perspective of domain bans being the only real scalable solution.

Maybe their "bring your own algorithm" idea will help, maybe it won't. Only time will tell, but I'm not holding my breath.

Mastodon

Yeah, know how I mentioned Mastodon as a project earlier in the article? They're trying to become a product. One of the ways they are trying to do this is by making mastodon.social the default homeserver when you set up the app. Again, can't blame them there. The most common bit of end-user feedback is that they don't know what a server is. Again, understandable. Individual servers being relevant is a protocol thing, and many users come from the perspective of using products.

That and "if I can move my account why can't I move my posts" that is. Both are incredibly valid questions and they really do need a good answer if the project wants to really graduate into a full-fledged product.

There's another bit about product development that I haven't really touched upon until now: products have a core philosophy to them. They are really a set of ideas that guide people building all of the projects that build up to the product. These things surface in the projects, but at some level they are a completely different thing.

I don't know if Mastodon can really become a product because of how diverse the community is. There's not really one common tie that binds everyone together, there are a lot of smaller communities that all interoperate.

In terms of mastodon.social, that may be able to become a product with one of the projects powering it being the Mastodon codebase, but Mastodon itself probably is never going to be a product.

At least in the fediverse nazi punks get blocked quickly and people crowdsource warning other instances about them. From empirical evidence, it's taken several days of consecutive outrage on Bluesky to get notable transphobes banned.

Every community being its own fiefdom really does make a lot more sense in the long run because that's just fundamentally how people gather into groups.

The path forward

I think the real solution to the protocol vs project problem is ultimately going to come down to communities banding together to create services to meet the needs of their members. Communities are what really drive society forward, not individuals.

Many people take this product vs project dichotomy to be made better by "personal responsibility" or something along that line, but honestly I think that having things like social media platforms being run by communities is a fundamental strength, not a weakness. Content moderation and server administration are fundamentally separate actions and things benefit from separate people having those responsibilities. The main problem is that a lot of these skills that were essential to early internet communities are locked away and relegated to experts that are paid too much to want to do this shit on their offtime. Can't blame them.

My therapist has told me that I really shouldn't be so conspiratorial-minded but goddamn this entire thing looks like there is some kind of conspiracy to make building communities harder and harder. Many online communities today barely resemble the communities that our parents and grandparents grew up with. Everything is so alienating and small.

This entire situation sucks and I don't know what to do about it. One of the main reasons I made my blog on my own domain hosted via fly.io is because I wanted to see how hard it would be to use Fresh to remake my blog in my image. Turns out it wasn't hard for me, but I'm an expert in backend software development and I have opinions about this shit.

I really wish I had answers here. I don't. However I think it's important for us all to really realize that projects and products are vastly different things. More than likely, you work on a product at your dayjob and work on projects on your offtime.