My least-favorite phrase to hear in a design meeting is "why didn't you just." It's an extremely efficient phrase--it packs both a dismissal of a specific idea and an insinuation of more generalized incompetence into just four words, without being overtly offensive. I hate it. So it might be surprising when I say that there is a "why didn't you just" at the very heart of my project now. I've had, if not a front-row seat to the recent seasons of Extractive Corporatist Theater, then at least a box with an unobstructed view and the potential for an invitation to the afterparty. And a lot of what's gone through my head, looking at--well, a lot of things, but social media in particular--has been: these services aren't good, and no one really seems to like them. Why don't people just notice that they're only using a couple of really inexpensive services and buy those services directly?

One self-interested reason to avoid saying "why didn't you just" is because one often looks like an idiot when the reason is supplied. And after about 7 months of working, as a competent-but-mediocre practitioner, on finding a way to buy "just" the services that people actually use in social media--let's say that I understand why the average non-practitioner doesn't "just" set this up for herself one night after dinner.

That said, as an answer to the question, "why don't people just operate their own social media," the response "because it's honestly too complex to set up" is kinda the best answer that someone like me could hope for. If the answer was, "because people like being surveiled and exploited," or "it's too expensive for the average person" or "it's not technically possible to buy these services in small quantities," then that might be a conversation-ender. But complexity is a matter of packaging. A TV is complex. A cell phone is complex. The equipment required to produce a salable bag of potato chips is complex. Nevertheless, these artifacts--by evolving into what I would call almond-shaped systems, have made themselves broadly accessible and commercially successful.

So why hasn't this happened for hosted internet services like social media? There are three cultural reasons that I can see. First, infrastructure is an afterthought in many software-writing environments. It is often, along with security and QA, considered a "support role" without much visibility or social cachet. There is a sense that if you are too focused on infrastructure then you must not be building the useful things^[1], so within software-writing as a practice, infrastructure tends to be overlooked. The second reason is that "giving your customers something that they can own forever" is widely recognized as a terrible business strategy--one example of this is how the Singer sewing machine company had a policy of destroying old machines that it received as trade-ins for new ones^[2]. The third reason is the distastefulness, to people who might otherwise be interested, of using the major cloud vendors, especially Amazon and Google, as suppliers. When you take these things together, you get a situation where individual practitioners, motivated by creativity and interest, see this kind of solution as "enterprise-y" and unsexy; where businesses, motivated by profit, see it as pointless or suicidal; and where people concerned with social justice won't buy the argument that, as a pragmatic move, using these services directly would vastly improve their negotiating power against the likes of facebook, whose entire business model depends on a captive audience not considering alternatives. And I don't mean to say that these are unreasonable positions for any of those actors to take--as a practitioner, a person who understands both economics and sustainability, and a person who cares about social justice, I can see how each social position--practitioners, businesses, activists--is drawn to its particular conclusion. I can't argue with the conclusions within those frames; all I can do is suggest that those frames are too limited for a solution to emerge solely within any of them.

I think of this as a problem of "perceived bigness." From an individual practitioner point of view, industrial-scale cloud services, however cheap they may be in operation^[3], are just too big and unwieldy to want to use. For a commercial enterprise, the idea of doing a bunch of work that is designed to leave you without a guaranteed revenue stream is insurmountable. For activists, the moral hazard of using massive corporations to oppose exploitative capitalism is too great to overcome, despite "using the adversary's size against them" being the classic successful strategy of effective asymmetric resistance movements.

But what if the kernel at the heart of all of this is not big, but small? What if the thing that makes these systems too big and enterprise-y for individual practitioners is demonstrably a matter of perspective^[4]? What if most of the complexity that makes facebook a behemoth is not the complexity of services offered, but rather the complexity of maintaining exploitative revenue streams, which we as users don't need or find useful? What if the moral concession, in terms of using huge cloud providers directly, is demonstrably much less than the concession involved in allowing oneself to be monetized by those same corporations in the status quo^[5]? If those things are the case, and self-run services could be higher quality and more efficient than centralized corporate services, then what we really need is not to meet the incumbents in pitched battle on their turf, but rather to provide an easy path for their captive audiences to escape, like drilling a small hole at the bottom of a dam, and letting time and pressure take care of the rest.