A Cornell Box I Made Out Of Algorithms
I’ll keep things brief this week as my brain is all filled up with reading our manuscript and I already begged our lovely whip-cracker crackerjack editorial shepherds with a three day extension to final-final stuff and they’d probably look at me with wonderment that I wrote 3000 words for a newsletter but somehow couldn’t find time to read and wordsmith the The Manual of Design Fiction. (There. I cleverly and surreptitiously settled the ‘what’s the title’ discussion.) Definitely a sideways-eye-narrowing emoji should go here, on behalf of my friends and colleagues working on the project.
This week’s newsletter is brought to you by “A Bestiary of the Anthropocene”, the latest work to effervesce from the bubbly brain of Nicolas Nova collaborating with a remarkable host of others (Anna Tsing, ftwi!) You’ll want this one, properly. No e-book bullshit. It’s an object, not a string of data. (But check out the free excerpt from the Laboratory Shop.)
But here’s what I’m thinking about right now which was prompted by that one bit in Nick’s awesome Ikea Futures Day presentation.
So — back to the Design Fiction archetype — a key element in representing possible future outcome. It is critical in setting context. In some cases, I’ve found that it is what kicks everything off, even before you’ve done a ton of background work to know what sort of problem you may be working on. The archetype is an inhabitant of the little corner of the world you’re investigating. It’s that scrap of ephemera, the found object, the MacGuffin that our time-traveling anthropologist brought back from their sketchy traverse because they could only grab shards and discards. The archetype has to tell a world of stories while we’re doing the work of imagining possible outcomes.
Parenthetically, the Annual Report I reported on last week? It was crazy how landing on that particular archetype to represent a future “as if” that future contained an enterprise that wrote that particular Annual Report led to months of deep fun, and engaged work. All in the service of thinking and creating through the problem to tunnel towards an outcome (one could MadLib the word ‘solution’ there) that answered the question: what does the future of the company look like? That way, with that outcome-solution, I could answer that question with substance and acuity. It may be the case that what I actually wanted was a 12 word pitch but, like I’ve said before — I went to Montessori School. So, like..
The Cornell Box is a kind of artificial, algorithmic test for establishing the apparent accuracy of a computer graphics rendering. It’s an algorithm for testing algorithms. Doing this was critical back in the Jurassic era of CG, when CG was a highly bespoke and cobbled-together affair. We’re talking about the late 1980s, early 1990s when you might build a computer from parts you got at a weekend Amateur Radio Enthusiasts’ Swap Meet. Lots of taut fellers with lightly soliled and puckered dress shirts with a missing button right at the most strained section, just below mid-chest. Glasses taped at the rim. (I know. I was there, too.) Now you can get commodity computer graphics from Amazon delivered early yesterday that can be used to draw photoreal landscapes while they mine bitcoin. But, back in those days, CG was enlightened, mesmerizing, jaw-dropping stuff. And it was just tea kettles and billiard balls reflecting an invisible light. Standing ovations were offered for having these hum-drum objects sitting on a reflective marble-like countertop. That’s all one needed to impress and amaze.
I happened to be in Ithaca New York at Cornell University when this was happening there. The high priests of the field were doing their whimsical magic right there in a new building adjacent to the old, crappy Electrical Engineering buiding I was relegated to. I think they had magstripe card keys to get in the new building lest some miscreant undergrad go lurking about. I think they got a new building from Sun Microsystems, if memory serves. And all the Engineering School professors got Sun workstations — the pizza box ones. And they were doing super high-res graphics on them cooked out by computers. It was amazing. (Some were satisfied with 2D bar graphs of experimental data, but they probably already had tenure.)
This was the highspeed stuff. Not the blocky bitmapped stuff from video games and ye olde Apple II. Well, it wasn’t real-time rendering but — damn — that box sitting in that Cornell Box looked like, well — what you might imagine it would look like if it was a box sitting in a fictional and perfect box lit from the top by a singe light source.
There was never the question as to why it was being done because it was so seductive and alluring, these renderings of boxes in a box. One simply wouldn’t ask “why?” — you just wiped the drool away from the corner of your gaping pie hole.
The Cornell Box at the time was a kind of fictional jig only meant, as near as I can tell, to isolate and constrain light sources so that one can create, study and improve upon the various ray-tracing and other rendering algorithms that were being crafted at the time. Nothing gets into or out of the Cornell Box. No light. No culture. No dust. No crappy instruction manuals. No Terms & Conditions statements or credit card scams. They have no context or utility other than abstracting away the rest of the world. One could hopelessly wait for the camera to pull back from the box so you can see the actual box on a table on a lab bench next to an oscilloscope, extensive and tangled ribbon cables leading to a rack of bleeping measuring instruments, leaky couplings on the back spilling some kinda viscous ooze. A styrofoam coffee cup nearby woud signal desperate humanity. None of that ever appeared. Frankly, I’d be amused and surprised, more as a cultural historian than an engineer, if the Cornell Box were still used in any way other than irony.
The Cornell Box and the style of rendering things in isolation, floating in a light gray environment, has been the de facto mechanic for showing future products in alluring and seductive ways. Outside of the world, quite literally inside a rendering algorithm that knows nothing about dings and dents, nor dusty grimey convenience store shelves where all great futuristic inventions and products go to eek out the ends of their lives, like some kinda commodity product eldercare facility.
When one holds forth on, say, ‘Workplace Futures’, or ‘Driverless Car Futures’, or ‘The Future of AI’ it feels to me that one would want some kind of ‘hooks’ that land in the experiential, lived aspects of a more embodied moment in these worlds. Rather than the Cornell Box of ‘AI Future’ — the bland, sterile, isolated, abstract consideration or speculation about what that is — Design Fiction goes for that moment in the life of someone who is living at that moment with AI. Like going to the pharmacy to get another month’s supply of Rejuvinex® — inload 12TB with breakfast, and wait 20 minutes before conversing with anyone but immediate family members.
Design Fiction is the world ‘as if..’ without even bothering to utter the word ‘future.’ Design Fiction is simply in the world and as futuristic feeling as a corner bodega laser pointer meant to be used as a cat toy, or a wrist watch that can take phone calls and measure and record your entire day’s heart rhythms. Which is to say that Design Fiction futures are meant to feel banal and routine and not the mouth-agape wonderment we typically associate with futurity. This is what makes good, engaging, thought-provoking Design Fiction. What my chum and fellow fellow Nick “Hallo! I’m Nick!” Foster refers to astutely as ‘Mundane Futures’. The future as normal, ordinary, everyday.
So I guess my point is that good, utilitarian, productive, evocative, conversation-worthy Design Fictions avoid the Cornell Box or the floating abstract gray environment and become part of the context of a lived world. Design Fictions come a bit scratched and something is chipped on the side and, dear me — hasn’t been cleaned in awhile and that’d be a tomato sauce stain right there kinda futures. Design Fiction is of the world, embodied in it Where the Action Is, to tip the hat to our friend Paul Dourish, who has what remains in my mind as one of the canonical book-length insights on embodiment as it pertains to humans interacting with algorithms.
Things in context tell stories about the world you are imagining. Things in a Cornell Box are just things looking pretty in a box.
p.s. With a tattered, ill-fitting Cornell t-shirt on, I solemnly salute the work the Cornell Box and Professor Greenberg and his colleagues have contributed just to be clear. It’s an object lesson I’m after here — not a rant about the Cornell Box itself. (And-also, you know -‘Go Big Red!’)