A Blog by Jonathan Low


Nov 25, 2016

This $1,500 Toaster Oven Personifies Everything Wrong With Silicon Valley Design

Solving a shrinking set of first world problems for an even smaller set of potential customers. JL

Mark Wilson reports in Fast Company:

Instead of teaching ourselves to cook, we're teaching a machine to cook.
I slide a piece of salmon into the June, one of the most advanced ovens ever built. Loaded with a camera, temperature probe, Wi-Fi, and algorithms, it'll cost you $1,500. It required nearly $30 million in venture capital to create. It was the brainchild of the engineer who brought us the iPhone’s camera and Ammunition, the design firm that gave us Beats headphones.
"We take very hard technologies, AI, deep learning, and lots of sensors, and we apply that to creating a well thought through, simple interface that just makes your life better," says June cofounder Matt Van Horn, another Apple alum who cofounded Zimride, today known as Lyft. "Our MO is we just want to inspire people to cook more." It's a tall order, but one that Van Horn delivers earnestly, the idea being that if cooking required less of us, we'd simply do it more. Yet in buying into the June, the home cook is becoming a consumer rather than a creator. The June asks cooks to put their faith in the fledging startup’s proprietary software getting better, rather than improving their own analog skills—skills that will work on any machine, in any kitchen.
Door closed, the oven knows it’s salmon. I press "salmon," and the June glows like a space heater, convection fans whirring. In precisely 10 minutes and 38 seconds my salmon will be done, the screen claims. Which seems way too fast, but what do I know? But 10 minutes becomes 20, and 20 fades into 40. It's almost seven, and still the timer's ETA is jumping around. This was all a replay of the night before, when our steak was cooking, and the June was texting messages like "NOTIFICATION_ETA_PESSIMISTIC"—a bug that the company would like to clarify it has since rectified.
The salmon's done at 6:52 p.m., when we’ve already devoured the sides that I'd rushed to assemble in my real oven, since the June only ships with a single rack.Update: June audited my oven’s data and claims that the salmon finished 31 minutes into cooking—still 3x June’s original estimate—and I left it in longer by mistake. My iOS logs show I did receive a push notification at that time, but that doesn’t account for what the oven itself was conveying. For me, being in the kitchen during this time, it’s impossible to know if this was all my error, or the result of what the oven interface was conveying over that time. More on that below.
[Image: courtesy the author]
"[The] salmon’s incredible," Van Horn had bragged earlier. Which seemed a stretch to me: "The salmon’s incredible" is what a waiter tells you when somebody at your table can’t eat gluten. Objectively, the fish was cooked to temperature and still moist enough—which you could have done in any oven, really.
This salmon had become more distracting to babysit than if I’d just cooked it on my own. This salmon had become a metaphor for Silicon Valley itself. Automated yet distracting. Boastful yet mediocre. Confident yet wrong. Most of all, the June is a product built less for you, the user, and more for its own ever-impending perfection as a platform. When you cook salmon wrong, you learn about cooking it right. When the June cooks salmon wrong, its findings are uploaded, aggregated, and averaged into a June database that you hope will allow all June ovens to get it right the next time. Good thing the firmware updates are installed automatically.
The promise of June is quite alluring. It’s simply an oven that senses progress in what it’s cooking via a thermometer probe. Its camera can use object recognition to automatically identify 20 different items—and eventually, the camera will be able to tell if a sweet potato is whole or diced, matching up the perfect, auto-cook recipe as a result. You can "set it and forget it," like Ron Popeil once said. There's finally enough technology to make that possible.
[Image: courtesy the author]
But the June's fussy interface is archetypal Silicon Valley solutionism. Most kitchen appliances are literally one button from their intended function. When you twist the knob of your stove, it fires up. Hit "pulse" on a food processor and it chops. The objects are simple, because the knowledge to use them correctly lives in the user. If you get the oven temperature wrong, or the blend speed off, you simply turn it off and try again. The June attempts to eliminate what you have to know, by adding prompts and options and UI feedback. Slide in a piece of bread to make toast. Would you like your toast extra light, light, medium, or dark? Then you get an instruction: "Toast bread on middle rack." But where there once was just an on button, you now get a blur of uncertainty: How much am I in control? How much can I expect from the oven? I once sat watching the screen for two minutes, confused as to why my toast wasn’t being made. Little did I realize, there’s a checkmark I had to press—the computer equivalent of "Are you sure you want to delete these photos?"—before browning some bread.
The June app itself, on your phone, is similarly loaded with information that reads like a red herring for real insight. It's content for content's sake. Aside from the litany of push notifications, which makes each dinner feel like a Tinder date who is VERY EXCITED TO SEE YOU AGAIN, you can see a schematic of just what June oven features (broiler, fans, etc.) are on, along with a live video feed of your food cooking. This latter feature seemed intriguing, since once something’s finished, you can watch a time-lapse of the food cooking—like one of those Pillsbury cookie commercials. Unfortunately, the playback looked something more like my roasted cauliflower’s sex tape. The unappetizing green glow that made me wonder, why would anyone bother to film this? The coup de grace of all the info-overload absurdity is a line graph, which compares oven temperature to item temperature. In case you weren’t sure why your salmon took 45 minutes, here you go. It’s science.
None of June’s mistakes are sins though. Isn’t the promise of mindless dinner on your table something that all of us—even people who love to cook—would agree is part of the TV-dinner-eating, slow-cooker-pot-roasting American dream itself? "We love cooking. We’re not trying to take anything away from cooking. Especially in 2016, we spend so much time at desks, looking at screens, it’s so much fun to chop carrots, garlic, preparing something, and creating a piece of art. We don't help you there," says Van Horn. "We help with, am I potentially undercooking this chicken and getting my family sick?"
And yet, June is taking something important away from the cooking process: the home cook’s ability to observe and learn. The sizzle of a steak on a pan will tell you if it’s hot enough. The smell will tell you when it starts to brown. These are soft skills that we gain through practice over time. June eliminates this self-education. Instead of teaching ourselves to cook, we’re teaching a machine to cook. And while that might make a product more valuable in the long term for a greater number of users, it’s inherently less valuable to us as individuals, if for no other reason than that even in the best-case scenarios of machine learning, we all have individual tastes. And what averages out across millions of people may end up tasting pretty . . . average.
Cooking has always been a highly personal, multi-sensory experience, where trial and error is the only way to become the all-star cook most of us know as grandma. But as I put the salmon on the table 40 minutes later than projected, I had no idea what I should have done differently, other than to never have used June in the first place.


Post a Comment