A new app unveiled at the ReWork Deep Learning summit last May 26 by Google's Kevin Murphy is the future of digital calorie counting. The technology comes from DeepMind, an AI starter that Google acquired in 2014 for the reported amount of $400 million. The app was immediately marred by online media as nothing more than a faulty and based on "bogus science", despite its obvious potential.

The app, named Im2Calorie, works by recognizing the photo you post on your instagram, measuring its size in absolute terms or relative to the plate it's served in. The program will then search its database and come up with a calorie count. Online publications jump at the chance to criticize its accuracy, citing the study published in the New England Journal of Medicine in 2011. Murphy claims, however, that the inaccuracies will be "minimized over time." Skeptics also mention the fact that the food labels the app gets its data from is also inaccurate. So what possible benefit can this app bring that Google actually patented it?

The Im2Calorie may not be the app of choice for the health conscious food photographer, but the story doesn't end there. If this app can detect food, it can be used to detect something else in the future. It can lay the foundation, so to speak, for bigger and better things. Imagine an app that has a huge database of visual information that can be translated into a recognition software of sorts. Looking back, we have similar apps that are tailored to specific niches; ASAP54 lets you detect clothes and shoe brands and last year's controversial facial recognition app, Name Tag, uses various criminal databases.

In theory, Im2Calorie is suppose to make you think twice about what you eat, despite its inaccuracies. The app's potential for improvement is great and we look forward to DeepMind's future projects.