The first phase was fitness trackers — devices meant to track how much you move. The idea is simple: by monitoring how little you move and telling you about it, a device can inspire you to move more. Simple. Except that many people who buy these devices don’t use them — 42 percent of the people who start using fitness trackers quit using them within the first six months, according to an NPD report.

That hasn’t stopped device makers from trying to move further into tracking behavior with calorie counters. This involves bad science, for reasons His Vladjesty outlined last year: calorie counting is ineffective and gives consumers a false sense of precision in what they’re eating. Even Weight Watchers, the pioneer of calorie counting, has abandoned it. But that message hasn’t made it to the enterprising entrepreneurs of Fitly, who are now trying to crowdfund a plate that weighs your food, pulls calorie and nutrient counts from the USDA, and adds them to a tracker.

I’m not going to rehash the scientific reasons the SmartPlate’s calorie-tracking focus is a bad idea — read Vlad’s post for that. Instead, I’d like to talk about how people actually eat. For starters, we don’t bring our own plates to restaurants.

The plate is part of Fitly’s goal to build "an ecosystem," Ortiz says. He tells me the important goal is personal accountability — keeping people aware of what they’re eating. The hope, he says, is that once people recognize their bad habits (in his case, eating his girlfriend’s leftovers), they’ll change. "Awareness drives change," he tells me.

I am on a somewhat restricted diet myself — I’ve been a vegetarian for almost 20 years now — and even that can be isolating: I have more than once been invited to friends’ houses for dinner only to discover that there is nothing I can eat. Well-meaning PR people have hoped to impress me by taking me to a steakhouse, where they find out that I’m interested in salads and Manhattans, not dead cows. Even as someone who’s accustomed to having to explain my weird diet, I cannot imagine bringing my own plate to a restaurant or dinner party.

The social power of a shared meal is exactly what devices like Fitly’s SmartPlate miss. Diets are already isolating ("No, Grandma, I can’t have the cheesecake you made specifically for this party, I’m dieting! What do you mean, your feelings are hurt?"). With the SmartPlate, whenever you’re invited to a restaurant or dinner party, you have to make a decision: do you want to bring your plate to a meal and risk ridicule, do you want to stay home, or do you want to abandon the plate entirely and eat like a normal person? It’s easier to stick to your diet if you stay home, of course — but that comes at a social cost.

Social costs shouldn’t be dismissed lightly. Take teetotalers, for example. Even if you exclude people who are in recovery, non-drinkers tend to die sooner than moderate and heavy-drinking counterparts and are more likely to be depressed. They also make less money. These findings are puzzling — right up until you remember how close you may feel with a colleague you killed a bottle of Four Roses with. We’re social animals, after all — bonding activities are important for our well-being.

The problems with Fitly’s SmartPlate don’t end there. It’s meant to be a home device, one for everyday use. At the demo I saw, the prototype plate wasn’t ready to connect to the USDA; rather, a much thicker cutaway version was used to demonstrate how the technology would work. Ortiz used the prototype to show me how accurate the weight sensors are, with two apples. The first, smaller apple, the plate clocked at 5.7 oz; a food scale showed that its actual weight was 5.708 — not bad. The second apple in the demo weighed in at 11.85 oz on the plate and 11.95 oz on the scale. Better than trying to guess by eyeballing how much the apple weighs, I suppose — but not quite the level of precision one would hope for, even if calorie counting did work. Ortiz used his phone’s camera to take photos of the apples, which the app readily identified. In the planned version of the plate, the cameras will be embedded, he says.

Some Brussels sprouts sat nearby, but we did not measure them. Instead, we moved onto bread. A few slices of white bread — one slice is apparently too light for the sensors to notice — were piled up next, then some slices of whole wheat bread. Ortiz asked me if I noticed any differences between the crusts and — to his evident disappointment — I said yes. There were admittedly subtle differences in texture and color, and I suppose I was meant to be impressed when the camera noticed them as well.

The SmartPlate is meant solely to count nutrients, but it struggles even there. That’s because the method of cooking you choose for dinner can shift which nutrients are present. Even assuming those nutrients are measured accurately, though, the way people absorb food varies, based on a number of factors — from what else is in the meal to which microbes inhabit the eater’s guts.