Windows for the Astronauts, On/Off Switches for Us
On comfort and control, or an antidote to automatism.
This post references the short story “There Will Come Soft Rains” by Ray Bradbury. Consider this a spoiler alert. Copies of the approx. five-page story are pretty easy to find, if you haven’t read it and want to before scrolling through this.
~1,000 words, a five-minute read
Before he demolished Ray Bradbury’s former home, purchaser and architect Thom Mayne justified his decision to do so in part by saying the building “wasn’t one of the more remarkable houses in the neighborhood.” Some pictures published in The Guardian when the property went on the market show that, indeed, it was a relatively unassuming yellow-stucco four-bedroom.
The fictional house dreamed up by Bradbury in his 1950 short story “There Will Come Soft Rains” was far more extraordinary, but it too would eventually be destroyed. The building’s demise is probably the second-most devastating part of the story, which pretty much has only domestic appliances for characters. Devastating part #1 is that the former human inhabitants have been recently vaporized in a nuclear explosion; their silhouettes imprinted on the wall a final snapshot of the family enjoying their yard the instant before the blast. Under a radioactive backdrop, the story turns to the robotic activities and “monologues” of the various gadgets that continue to operate like normal in the aftermath of the nuclear apocalypse.
Bradbury’s fictional household gizmos have the unmistakable aura of mid-century American mechanicalism. They make and clean up after a full breakfast, light cigars, set up card tables, ask wandering wildlife what the password is, and dispose of deceased pets. Some rhyme. One recites a bedtime poem for the late Mrs. McClellan, her apparent absence hardly a barrier to the reading. Overall, Bradbury evokes a domestic scene where every thinkable human task is attended to at the appointed hour by some quirky, space-age sci-fi doodad.
And ever since they replaced the light switches with automated sensors in the undergrad architecture studio from whence I came, I think of this story from time to time. Allow me to list some examples that induce the recollection to percolate by way of a bullet pointed list:
In the aforementioned undergrad studio when we wished to keep the lights off in order to test a lighting mock-up, or (more often) desired an ambience with only task lighting, but always loomed a hopeless inability to do this after 9pm with any other method besides hour-long stillness.
In my experience the water-detecting windshield wipers in newer cars have a very spooky and irregular cadence and the regular modes, while still available, are much preferable.
The toilet that automatically flushes too much and at the detection of the slightest movement.
A toilet with no button at all and only motion sensors, and when those fail, eeek!
An office space where the air conditioner runs only at regular hours but inhabitants work irregular hours, often leaving the space cooled and empty or sweaty and occupied.
But none of that was as disturbing as what happened when I recently walked into this terrace area:
I won’t say where it is — that’s irrelevant for the purposes of this post — but in the context of the other issues listed, it may take the prize of being the most blatantly problematic from a design perspective.
The sequence is as follows. When I walk in, a motion sensor beeps a high b-flat. The fans turn on. It’s a welcome draft in a very warm climate; the fans themselves are clearly a good idea. I sit down to type and read. Five minutes later, the fans turn off. The beep was from a motion sensor that detects movement, and typing isn’t enough to trip it. Neither, apparently, is waving my arms about. The motion sensors are mostly pointed toward the adjoining corridor, not the seating area itself. To be continuously fanned in this space requires one to waff a giant stick (or something), or to get out of the chair and walk about every five minutes.
Yes I know incremental improvements could solve some of this. The timer could be timed longer, and the motion sensor could sense more. But pretty soon we’re solving problems with problems, and anyway this example is merely a way to raise the bigger issue, which I will introduce by asking: what could we call this troublesome over-reliance on sensors that become the sole arbiters of operance?
“Automation” is an easy first choice, but I’m tempted to go for a rarer, less optimistic, and more philosophical term. That would be “automatism”, the OED definitions of which is a little scary when considered seriously:
The theory, belief, or doctrine that living organisms act purely mechanically, like automata, and are motivated by physical causes, rather than consciousness, intelligence, or will. Also: the condition or state of being so motivated. Now chiefly historical.
I’m not saying some sort of total-automatism state is arriving and/or inevitable, nor will I even make the sophomoric claim that I made in an early draft of this piece and say “the world is increasingly automatized”. But, since the condition of total-automatism is arguably worth avoiding, let’s think about what it might actually entail as a quick thought exercise. To me, it gets into “you will own nothing and be happy” territory. It’s not exactly that. “You will control nothing and be comfortable” comes close, though. In fact, I’m not sure which of those is worse. Maybe they’re two sides of the same coin.
Luckily, total-automatism is stubbornly avoidable thanks to one weird trick: manual overrides. Otherwise known as the humble on/off switch. If that sounds anticlimactic, allow me to make a grand comparison to the astronauts. I insist that we maintain the provision of manual on/off switches in the same way that the astronauts insisted that spacecraft had windows. There are things we just need, despite the tendency of Engineering to optimize for performance and predictability, or however the algorithmic incarnation of the ideal is formulized.
With a switch we can turn off the lights, turn off the A/C, flush the toilets, and turn on the fan.
(And while we’re here, let’s make sure there’s an on/off button on the AI machine too.)
I finish proofreading, I get up and leave the terrace. As I walk out, the fans bleep on for the second and final time, earnestly breezing the now warm but empty chair. For the next five vacant minutes, the space is a living slice of Bradbury’s imaginary, post-human, gadget-charactered world. We all knew it would, but his fiction has officially outlived his house. Now we’re left to ponder what exactly might outlive us, and for how long. Terrace-wise, the fan’ll spin for a bonus five minutes.
Another example of automation and sensors gone amuk is the supermarket's self-checkout. Rare is the day that the sensors do not detect some error of your actions that requires intervention by a store employee. This morning at checkout, some detected offense required intervention by an employee then his manager, neither which could fix whatever incorrect action the sensor detected. Meanwhile, the ice cream was melting.
Ray Bradbury further explored the “empty house automata” theme in his Martian Chronicles: upping the Martian house tech with a sense of self preservation. Maybe next time that auto-fan will automatically clear the occupants from the room when the fan times out. A perfectly logical solution—in a mechanistic sense—to the author’s problem!