“Then she generated the light, and the sight of her room, flooded with radiance and studded with electric buttons, revived her.”
E. M. Forster, The Machine Stops
We have long dreamt of sentient homes and domestic automatons, with many such technologies actually coming to fruition in some form or another. Generally, they come as products which restrict customization, slouch intentionally toward obsolescence, and entail invasive telemetry practices as a crucial part of their business models. Does this promise of household intelligence, of an enlivened locale, truly require us to concede the privacy and autonomy of our own homes? Even putting aside the question of privacy, we still incur the risks of technical dependency on external parties, parties with policies and imperatives which often bear little obligation to adhere to the interests of those whom their technologies nominally serve. Indeed in this context we are framed as the users, our individual and communal nuance notwithstanding, and it is in this capacity that we receive such fickle fire.
Even in the language of institutional economic liberalism, technological progress can force paradigm shifts which, in turn, scramble the relations of production. That is, the nature of the tools we have at our disposal is liable to change, but naturally that depends on who is building what, and for whom. The future may be unevenly distributed, and capital may tend to accumulate upward, but there nonetheless remain windows of opportunity to seize, in the spirit of the public good.
In more direct terms, there are futures in which we can benefit from affordable “homecooked” technologies, technologies which recognize us not as impersonal users, but as people. This can allow us to enter into different relationships with our tools and our home, in greater confidence that our interests are represented in the very construction of these technologies. The lowering cost of software development, combined with the extensive availability of reliable open-source software, may succeed in rendering this alternative future more economically viable for wider and wider swathes of the population.
In our 20-person household called Embassy, which also operates as an active community space, we have home servers running run local-first apps and services. We run Home Assistant as a unified home automation framework, which can connect with our solar panels, our dryers, our light bulbs, our ESP sensors for temperature and humidity, and more. We host our own private AI, which we can use offline to chat with our finances, our appliances, and our knowledge base, without needing to trust any third parties. We have a media server which stores our film library, and speakers which automatically play classical music in the morning. We can download medical guides, appliance user manuals, and other key documents needed in the case of an emergency, so we can ask our house AI to explain these things to us where we lack the technical knowledge. All of this works offline: no subscriptions, no telemetry, no third parties involved.
We are also exploring various other components: self-hosted biometric wearables; NFC chips around the house to explain house protocols to guests; cryptographically secure tools for communicating complex private information between community members; a self-hosted website and wiki; robots which use our local AI to chat with people in the house; offline lending platforms so community members can browse and borrow from our book collection, as well as from our tool collection; and ESP sensors placed in plants to signal vegetation health; and local-first circuit-specific power monitoring.
Many of these components become more interesting as their utilities compound upon one another. For example, if we had self-hosted analytics for biometric rings, decibel sensors around the house, and speakers which can play white noise, then we can probabilistically modulate a white noise soundscape throughout the house to smooth out spikes in traffic noise during the night, based on readings of REM sleep disruption from our rings. If we had daylight sensors near our plants, those readings can be wired through Home Assistant to dynamically turn on grow lights near the plants; and likewise could soil moisture sensors be configured to activate automatic drip irrigators. If we had motion detection sensors throughout the house, we could program lights to turn off past a certain time of night, after a sufficient break in detected movement in the room. All of these things would operate in a unified system which would be legible to our local AI, and kept private from outside commercial interests.
Economically speaking, because there are 20 of us in this house, our collective benefit from these systems renders them more accessible then if we were to each live in our own solitary apartment. Many of the systems in question would qualify as club goods: non-rivalrous, but excludable. That is, their value isn’t exhaustible in nature, but access to them can be limited to those in the house or wider community.
While these systems don’t come with subscriptions or external dependencies, they are quite complex to configure. Even with the assistance of commercial cloud AI, like Claude and ChatGPT, there is still a considerable amount of technical expertise required to implement these things, expertise which not everyone has the time to cultivate. While it may just take one or two technicians in a community to manage these systems, most people do not live in such a community, with many lacking in-person access to such communities altogether. In this sense, many of the technical solutions described here are largely precluded by the wider status quo of loneliness and isolation. Lack of access to community, and social support networks they provide, deprives us of the opportunity to exercise relational skills. The deterioration of these skills tends to make it harder for us to relate to other people, thus further diminishing the presence of healthy community.
Guests who have stayed at our house, even for a few weeks, have said it has changed their lives, as if they’ve stepped into an alternate reality where our relational needs, hitherto starved, are met with abundance. Our technological solutions may help stitch the yawning interstices of our social fabric, but how we show up for ourselves and for others is arguably the behavioral basis for community, and by extension for democracy. What wonder is it, then, why our world is governed by grifters, when we ourselves fail to support those closest to us?
