Back when I was running Elgg, I'd meet someone every few weeks who wanted to build a competitor to Facebook. Inevitably, they would propose to do this by copying all of Facebook's features verbatim, but (for example) without an ad ecosystem or with a different algorithm for surfacing content. All of them were doomed to fail.

These days, I'm more distant from the alternative social networking ecosystem, but it's easy to spot the same ideas. One might propose a decentralized alternative to Facebook that has all of Facebook's features, for example, and assume that people will flock to it because it's not owned by a corporation. You care about privacy and ownership, after all - if others don't, surely it's just a matter of educating them?

Aside from with a relative handful of enthusiasts, these efforts are probably all doomed to fail, too.

The thing is, privacy and ownership are important, and over the last few years we've seen our quiet worries about silos of data owned by single-point-of-failure corporations grow into a global roar about their role in supporting pogroms and undermining democracies. Nonetheless, we've learned pretty conclusively that privacy and autonomy are not virtues for everyone - actually a lesson learned again and again in the 20th century in particular - so if we want these values to be adopted, we must find another way. The stakes around getting this right have never been higher. (It would have been nice to have gotten this right in 2015 or so, but here we are.)

People, in general, want convenience from their technology, not morality. So instead of building a more ethical version of the past, we need to build a more suitable version of the future. It turns out that data silos have left room for plenty of innovation here: how many people send emails to themselves to save a note, or have had trouble AirDropping to an Android phone? Why do I have to download WhatsApp to talk to my friends in the UK? There are lots of tiny inconveniences that would be made better with openness and a user-centered model.

The same is true of online communities. An artists' community has radically different needs to an activism community, yet on the silos they're shoehorned into the same interface and set of features. Communities for people with restricted vision or motion might perhaps be the most obvious example: why should they have to struggle to use interfaces designed for others? Or better put, why can't they have an internet experience designed for them? A federated galaxy of community platforms, tailored for the specific human communities that use them and linked by Google-like sites that facilitate discovery, would be a more functional internet for many people, and would also decentralize the social web. Over time, discovery could be decentralized, too.

Whatever we're building, we never absolve ourselves from the need to understand our users as people and meet their needs. We might have our own values that we want to convey - software as polemic - but we can't simply inject them into the status quo. We've got to use our values, our intuition, and our understanding of the people we're building our software for to build something new that serves its purpose better than anything that has come before it. That, and nothing less, is the job.