At the end of 2012, working hard on my own connected lighting startup, MooresCloud, I got very excited to find out that Philips planned to launch Hue, the company's own full-spectrum connected lights. I bought a ‘starter pack’ of three soon after release, and played with them for weeks.
It immediately became clear that Philips had put everything they understood about the psychological qualities of illumination into the Hue bulbs. Although they could produce a broad selection of reds, greens, blues and whites, almost all of the colours were highly desaturated pastels. Why, when equipped with the infinite variability of LEDs, would Philips limit their bulbs like this? Because fully saturated colours - particularly greens - make people look like zombies.
Philips had not marketed connected stage lighting; Hue were connected lights for the home. Its restricted colour palette made it nearly impossible for a connected lighting enthusiast to inadvertently create a horror show - saving users from themselves. That is, after all, what an expert should do.
The bulbs are so beautiful, their colours so rich, that I immediately put purchased a pair of table lamps, and put them at opposite ends of my office desk, pointed at the wall, where they throw a warm, buttery light that I find perfect for work. There they’ve stayed for almost three years.
But I need to be careful.
If those bulbs ever lose mains power they lose their minds. Or, specifically, they lose their settings. Plug them back into mains power, and they immediately burst into a fierce blue-white glow that makes me wince.
Why they should act in this way is beyond me, but one thing is not: this is exactly what Philips intended they should do. Every aspect of the Hue is so artfully designed, it’s inconceivable that this power-on behaviour wasn’t carefully engineered to deliver a particular experience. It’s just that in this case, this experience is entirely wrong, stupid, and annoying.
Most significantly, this behaviour means that the way we control lighting - switches - is suddenly useless. Every time a Hue bulb loses power, it loses everything that makes it interesting. I very quickly learned that I had to treat my Hue bulbs with caution, interacting with them only from the smartphone app.
That’s an unnecessary pain. It’s not as though these bulbs are stupid. Each of them have control electronics, and they talk to an ethernet connected ‘hub’ that provides a full REST API for monitoring and controlling every aspect of the bulb’s operation. Somewhere underneath all of this, there’s a lot of very sophisticated firmware.
Yet, somehow, Philips’ designers decided that it will not keep state over a power cycle.
Perhaps Philips’ designers really believe that a power-cycled bulb has to glow with full intensity when powered up, maybe in order to convince a naïve owner the bulb is actually working. Or perhaps they just hadn’t factored in a power switch.
My guess, three years on, is the latter. Lost in the capacities of connected lighting, Philips forgot how people use lights in real homes equipped with real lamp fixtures.
Besides, you can always reset your blazing lights back to comfortable levels through the smartphone app, right?
Yes, you certainly can. If have your smartphone at hand. Provided you can remember which folder the app's in, and which bulb is in which lamp fixture. And so on.
Every step along the path from the simplest way of making a device to function adds more ‘cognitive load’. We end up holding the bulbs ‘in mind’ because the designers at Philips felt that was a reasonable trade-off for the increased capacity of the Hue.
They may have had a point … back in 2012. At that time there were very few connected products on the market. Before I bought my LIFX bulb (which, shock horror, does remember its settings across a power cycle), my Belkin WEMO mains switches and so on, until - just last weekend, I purchased a Beurer connected scale, which dutifully records my weight, BMI and percentage body fat to my iPhone and HealthKit. (Presumably sending the data along to NSA, for all they’ll care about how my diet is going.)
I keep all these apps together in the same folder, and it’s getting a bit crowded in there. I resent opening yet another app to talk to another connected thing. Each of these apps have their own interfaces, which I’ve had to master. They have their own needs, which I must service. And they provide their own highly specialised data stream - which none of the other apps understand.
It’s bad enough today, but what happens when literally every appliance in the home - from the teakettle to the aircon to the oven - is connected? Never mind the huge security issues that open up - which I’ve already covered - how can we expect to juggle all of these interfaces and expectations?
Apple with HomeKit and Samsung with Artik and Google with Brillo promise one-size-fits-all solutions. Each will mandate hardware designers to bend their knee to their particular way of doing things. But even if they unify the interfaces, will this really make these devices easier to use? No one wants to have to activate an interface every time they use something. It’s an additional burden - and a mostly unnecessary one.
Why bother adding intelligence to something unless it’s capable of thinking for itself?
My Hue bulbs can think, “This is how we are normally set, so let’s go back to that setting after a power cycle." My washing machine can think, “This seems to be a load of sheets, so I’ll use my warmest settings." And so on.
None of this would require any action at all on our part, because devices are doing the thinking for themselves. No, they won’t always be right, and that’s when we reach out with those interfaces to amend those assumptions. But the connected world should be asking less of us - not more - doing more thinking, and leaving us to think deeper thoughts. Until we cross that Rubicon, consumers will reject these connected gadgets as too much work. ®