Blippar now recognizes cats, other items without AR codes
Augmented reality app Blippar is hardly new (it’s been around since 2012, almost as long the lab!), but it’s back in the bright lights after curiously shifting its focus to features associated with machine learning.
Until the latest update, the U.K.-based Blippar was mainly focused on advertising and content for brands. In a nutshell, Blippar allowed customers to “blipp” (or scan) a logo on its partners’ products to get more information. So, kinda like a QR code, with gusto: Blipps can generate a variety of actions on the partaker’s mobile device, like launching audio or video clips, dialing a phone number, or connecting to a social media page.
Blippar gained quite a bit of traction with its initial concept, partnering with several large brands and media companies including Conde Nast, TIME, and Unilever. But over the past year and a half, and for reasons unknown to us, founder Ambarish Mitra shifted Blippar’s focus towards machine learning, even going so far as to move his team from the U.K. to the Silicon Valley.
With the most recent update, Blippar gained the ability to recognize objects that don’t include a logo – such as dogs, apples, and cars. And as Mitra puts it, in an interview with Re/code,, the technology is still in the early stages, like the brain of a six-year-old. It can recognize “car,” but not “Prius;” an item of clothing, but not the label.
So naturally, we took it for a spin to see for ourselves how the technology works.
When I aimed my phone at Raya the cat, sitting in the windowsill, a flurry of words swirled around my screen like a tag cloud. And initially, the app seemed to have some difficulty distinguishing between species. It got “mammal” right away, but seemed to think my cat was a dog (to its credit, she does play fetch and growl at the mailman.) Finally it caught on and started bringing up cat-related words like “curiosity,” “kitten,” and “tabby.”
An icon of a cat appeared on the screen and when I tapped it, more icons appeared with different actions – giving me links to buy pet food, watch cat videos, and donate to the Humane Society.
But, in contrast to most apps that tie into e-commerce, Blippar wasn’t able to read barcodes when we tested it (however, it did spout a bunch of words like “savings” and “finance” when I aimed my phone at a barcode.)
Next, I stepped outside to go see what Blippar made of the chickens. As soon as the flock appeared in view, Blippar blurted another jumble of words, like “farm,” “food” and “nature.”
Impressively, the app even seemed able to distinguish between the hens (which it called “beautiful,” and we have to agree) and the rooster, zeroing in on Caesar’s tail with the word “cockerel”:
Back inside the house, it thought it saw a “squash racket” in the kitchen and initially mistook a pair of scissors for a table knife, but then quickly got it right. When I pointed my iPhone at an orange, it immediately displayed an icon of an orange followed by a bunch of juicy descriptors:
In general, the app did a remarkably good job of recognizing anything I put in front of my iPhone lens. Until I got back to my desk, that is. When I aimed my phone at the glass of water sitting next to my computer, Blippar declared it happy hour.
“Vodka!” “Cocktails!” Even whisky and wine – it must have chimed in with every type of alcohol (and juice) — but didn’t once mention water.
It’s Thursday morning, Blippar
Thinking the Blippar app might be confused by the fact that I was drinking water out of a pint glass with the words “hard cider” on it, I filled a plain water glass instead. But as I focused on the glass of water with my work still in the background, it did the same thing.
It’s only water. Honest!
Then a glass appeared on screen with a bunch of options, and I thought Blippar might have some sort of crowdsourcing feature that could help it learn the difference between water and whisky. Alas, no, but it did give me the opportunity to shop for water glasses online, learn about how to safely dispose of a broken glass, and watch a video of how to pour the perfect beer.
Aside from the fact that I felt a bit like the app was peer-pressuring me to pour a cocktail at the beginning of a workday, the revamped Blippar is surprisingly entertaining – particularly when you take it outside for a walk – and we’re interested to see how its machine learning evolves over time, and what can be done with it.
We can think of plenty of possible uses down the road (for example – describing the view through the lens to visually-impaired people). But in its current state, Blippar doesn’t provide much more than entertainment and a few links to buy things online. Is Blippar just a blip on the radar, or will we see it expand into something more useful? We’d love to hear your thoughts in the comments.
Starry-eyed surprise: The rise of wireless Internet startups?
By Leslie Ellis and Sara Dirkse
In case you’ve managed to forget the tragic tale of Aereo, brace for some flashbacks. Aereo founder Chet Kanojia announced last week that he is introducing a new Internet service – called “Starry” – that uses millimeter wave technology to wirelessly deliver Internet speeds up to 1GB. Kanojia aims to disrupt the business of broadband with this latest venture, a new take on fixed wireless technology.
“It costs the cable guys around $2,500 per home to deal with the construction costs of laying down cable,” said Kanojia on a Jan. 27 phone call with TechCrunch, setting the scene for his next big reveal. “And beyond cost, there are regulatory hurdles that slow down the process.”
We can’t really speak for the legal or regulatory challenges Starry might face (thank the heavens and – heh-heh! – stars), but it’s probably safe to assume that ISPs are already all over it. However, on the latest episode of Re/code Decode with Peter Kafka, Kanojia shared a more optimistic view. “Comcast isn’t going out of business,” he said when asked if he expected another legal challenge. “It is better for them to have a few small competitors in the market so that the government is satisfied.”
Potential legal battles aside, it appears Starry might face some challenges on the technical front too. Re/code’s Ina Fried put together a great analysis of some of the technical stumbling blocks that Starry might face, based on issues that plagued similar ventures in the past.
Here’s the basics: Starry uses technology called “millimeter wave band active phased array technology.” This being Translation Please, indulge us in doing what we endeavor to do.
Starry Beam rooftop module (Image credit: Engadget)
In a nutshell, Starry puts a bunch of little nodes called “Starry Beams” on rooftops in densely populated areas (this might ring a bell with anyone who used Aereo), and gives its customers little receivers called “Starry Points” to stick outside of a window.
Starry Point window receiver (Image credit: Engadget)
The Starry Beam shoots out millimeter waves in a bunch of different directions (this is what they mean by “active phased array”), and these waves bounce off of buildings and other obstacles until they reach your Starry Point. Aereo claims that they can transmit a reliable signal without line-of-sight from the node to the receiver, unlike previous attempts at fixed wireless Internet.
Aereo also offers the optional Starry Station, a $349 Android-powered router with a touchscreen that monitors its own connections, handles parental controls, and a bunch of other tricks.
The (optional) Starry Station (Image credit: Engadget)
According to Kanojia, this approach will save Starry, and its customers, a lot of time and money. “We can deliver faster broadband with no regulatory wait time and it will cost us only $25 per home,” he said on the TechCrunch call.
No word yet on pricing, other than that Starry will be a tiered service based on speed (topping out at 1 GB up and down).
Chet isn’t alone in the landscape of wireless startups targeting larger ISPs. Two San Francisco-based companies, Artemis Networks and Webpass, are working together on a similar service. This one is particularly interesting because Artemis claims to have invented a way to use interference as a channel for transmitting and receiving data – so the more interference, the better the connection. (It all sounds so … Hedy Lamarr!)
We can’t wait to see this one in action – and according to Artemis founder Steve Perlman (yes, that Steve Perlman — former head of WebTV, Rearden Steel and Moxi), we might see it in Boston later this year (assuming the FCC gives the nod).
Plus, Chet’s just an interesting guy to keep an eye on. He started out in cable, founding Navic Networks in the early 2000s. Remember? Microsoft bought Navic in 2008 for a reported $250 million; Chet’s been throwing wrenches at the industry that launched him ever since. Which makes us wonder: Who pissed Chet Kanojia off so bad?