Bonaverde wants to be the Keurig of raw coffee

Many people take their coffee drinking seriously, but even those with a proper espresso machine at home won’t roast their own beans. This delicate step is typically done on an industrial scale using specialist equipment, far from the end consumer. Now, the caffeine addicts at Bonaverde intend to bring roasting to the kitchen counter with the first all-in-one machine that turns raw, green coffee beans into a cup of joe. The company crowdfunded its appliance way back in 2013, and several years later the consumer model is now ready. But several questions have loomed over Bonaverde’s vision to change how people drink coffee — namely, where on earth do you buy unroasted beans?

Raw, green beans aren’t exactly easy to come by. They’re usually traded by the ton, not something you pick up at your local store. Some specialist retailers do sell raw beans in consumer-friendly quantities, but you still wouldn’t call them a particularly accessible commodity. Thus, Bonaverde hasn’t just spent the past four years refining its machine; it’s also established a complete supply chain to ship beans directly from coffee growers to your doorstep.

Bonaverde had to develop another machine to enable growers to package their product in the special parcels required by its all-in-one appliance. Like Keurig or Nespresso pods, these are not optional. The little pizza-slice-shaped bundles contain enough raw coffee for one pot, with the packaging doubling as the brewing filter (that’s one less thing to buy). The grower-side equipment is built with the isolated fields of developing countries in mind. It runs on solar power with battery backup, is simple enough to be fixed with “a hammer and screwdriver,” rather than needing elaborate replacement parts when it breaks down, and even has a few bonus features like being a WiFi hotspot (an embedded cellular SIM handles the connection).

Key to the packaging process is a unique RFID chip that’s attached to every pouch. Bonaverde users will have to tap this tag on the front of their machine before it’ll roast raw beans. This is partly for quality control, the company tells me. As the end user will likely have zero experience with roasting, the tag communicates the optimal profile to the appliance — different strains of bean from different regions require distinct temperatures, timing and air-circulation settings to be at their best. In other words, Bonaverde doesn’t want consumers to have at every parameter, ruin the roast and think their machine is junk as a result.

End users do have some level of control, though. Using a companion mobile app, they’re able to tweak the roasting profile if they want a more intense flavor, for example, and they can also determine how finely the roasted beans are then ground, which impacts the final brew. (The idea is that this take on personalization is better than asking users what exact temperature they want during the final phase of roasting.) Bonaverde plans to introduce community features to its app in the future, too, so connoisseurs can share their tweaks with other users. Importantly, the app also has a scheduling feature so your pot will be ready at a certain time, meaning you don’t have wait for the roughly 20-minute roasting, grinding and brewing cycle to complete each morning.

The RFID packaging is also important to Bonaverde for a completely different reason. Being a startup, it doesn’t have the cash to set itself up as a middleman between grower and consumer, buying from one and selling to the other. Instead, growers effectively license Bonaverde’s vision, packing the coffee off their own backs. Bonaverde handles the shipping — a more manageable overhead — and distributes the coffee to consumers, who just pay for delivery. Buying the actual coffee occurs only once you’ve tapped that RFID chip on the appliance, at which point the grower gets what it’s owed and Bonaverde takes a cut. In this way, the startup hopes to build a global ecosystem around home roasting without becoming a trader itself, thanks to the microtransaction model. (You could just… never actually scan the chip.)

That sounds potentially exploitative and risky for the grower, but Bonaverde is guaranteeing its partners a minimum of 30 percent more than Fair Trade wholesale price with its model. The farmer is able to set their own pouch prices, which can range from roughly $2 to $5 per pot — handpicked beans that are more expensive to produce could demand a higher price, for instance.

In some ways, Bonaverde has it all sewn up. Its all-in-one coffeemaker wasn’t going to be viable without easy access to raw, green beans, so it’s creating the supply chain. The startup doesn’t want to own the whole ecosystem, though, and hopes to partner with companies as well as growers, so you might pick up compatible pouches at your favorite local spot. Familiarity for the customer, branding for the coffeehouse, and a cut for Bonaverde. Putting pouches in coffee shops and machines in hotels is likely a ways off, though, as you can’t solve the inherent chicken-and-egg problem overnight.

Bonaverde has a long road to becoming the next Keurig or Nespresso. The obvious obstacle is growing the initial user base, thus becoming more attractive to pouch partners, which in turn makes the machine itself an easier sell. But the company is asking for a significant amount of commitment. First, there’s the price of the all-in-one machine.

While some early-bird backers managed to pick one up for as little as $250 four years ago — and are just now expecting delivery — the appliance goes on general sale today for $799/£799/€799 for the white version and $1000/£1000/€1000 for the silver model (coinciding with the launch of an equity crowdfunding campaign on Seedrs). New customers are looking at a fall shipping window, mind. It looks significantly different from the initial Kickstarter pitch. It’s essentially a tall, square filter coffee machine with some added bulk on top to account for the roasting element. Not unattractive, but potentially too demanding for smaller kitchens. Beyond the machine, you have to be comfortable allying yourself with the whole ecosystem, from buying the pouches to the replacement filters that capture unwanted oils and such created during the roasting process (each is good for 30 brews, I’m told).

Ultimately, though, Bonaverde is attempting to change perceptions. The main selling point is truly fresh coffee anytime you want it. I tried a cup myself and found it to have a certain cleanliness about it, though it was also relatively tasteless. I’m pretty sure the company plied tasters with a light brew that was bound to be universally inoffensive, though, and I’m no coffee expert, so am hesitant to make any real judgments on quality, especially after just two mugfuls. Green beans apparently last upwards of a year before starting to go stale, and then there’s supporting the growers directly. But the freshest of coffee is still the primary draw.

I’m not sure real aficionados want to go to the trouble of roasting their own beans — this is also assuming they believe the machine’s process is as robust as industrial roasting — for the sake of a pot of filter coffee. Bonaverde intends to release an espresso model in the future, possibly later this year but likely next. This will at least give consumers options, but we’re still talking about an unexplored market here.

Bonaverde obviously believes there’s a large appetite for home roasting, and the company has reason to. The startup’s crowdfunding campaign eclipsed its initial $135,000 goal by over half a million bucks, and I’m told a pre-market competitor in San Francisco is currently working on a similar all-in-one machine. But whether there truly is a big enough demand for home roasting to support Bonaverde’s supply chain — or whether the company can create it — is the next looming question to be answered.

Source: Bonaverde

California may give EV buyers instant rebates

If you’re worried about the federal tax credit for electric car purchases phasing out, you may want to pay attention to what California is proposing. The state is concerned about demand for EVs declining as the tax incentives disappear. As a result, California is considering legislation that will give instant rebates to buyers of electric cars at the time of purchase.

A key component of this bill is that the rebate is instant, applied automatically when you purchase the vehicle. California currently has a mail-in EV rebate program for certain car models, but purchasers have to wait up to 90 days to receive their funds. It’s possible that this wait makes people less likely to choose electric over gasoline power for their vehicle.

While it’s not clear exactly how much each rebate would be, The California Electric Vehicle Initiative would be income based. The state is specifically concerned with helping low-income residents purchase EVs. Their goal is to bring the purchase price of electric cars down to their gasoline-powered equivalents. The bill currently proposes putting aside around $3 billion total for these incentives.

The legislation hasn’t been passed yet; while it passed an assembly vote, it still needs to face two Senate committees. It’s notable, though, because California is often a leader in emissions and car regulations for the rest of the country. If this becomes law in California, it’s possible we could see other states following suit.

Via: Bloomberg

Source: California State Legislature

Revamped Twitch app streams live from your phone

As useful as the Twitch mobile app can be, it still falls short of what you’re used to on the desktop, especially if you’re a broadcaster. Would you believe that you have to switch apps just to stream live? Thankfully, it’s shrinking that gap today. It’s starting to roll out a Twitch app overhaul that adds some of the features you take for granted while refining the interface. For creators, the biggest deal is simply direct mobile streaming — you can livestream to your channel without any go-between software. While phone-based broadcasting was certainly possible before, this could easily lead to more IRL streams from your favorite streamers.

Of course, there are plenty of updates for those content to sit back and watch. The Pulse discovery feed and the notification center finally made it to mobile, for starters. The interface should also make it easier to find those features, including navigation bars that promise easier access to mainstay features as well as swipe-based controls for picking videos or going back to a previous stream. You’ll also see a dark mode for those nighttime viewing sessions, and instant playlists that you can access just by swiping down.

You might not see all these features right away, as Twitch doesn’t expect everyone to get the new app until early July. When it arrives, though, it could shake up your expectations for Twitch. You’ll have more reason to watch on your phone instead of waiting until you get home, but it’ll also help Twitch in its quest to expand beyond game streams. It’s also good news for broadcasters who’d rather not be chained to their computers — they can go on a PC-free vacation without losing touch.

Source: Twitch

US hit by cyberattack that targeted Ukraine and Russia

Yesterday, a number of Ukrainian and Russian companies and state agencies reported being hit by a cyberattack, the results of which ranged from flight delays at Boryspil airport to a shutdown of Chernobyl nuclear power plant’s automatic radiation monitoring system. And while those two countries took the brunt of it, the virus at the root of the attack quickly spread throughout Europe and to Asia, Australia and the US.

Among those affected in the US were hospitals, the pharmaceutical company Merck, Nabisco and Oreo. A major Los Angeles port was forced to stop operations yesterday because of the attack and was still closed as of this morning. FedEx also experienced disruptions in its TNT Express delivery service. A US nuclear power plant was the victim of a cyberattack as well, but it’s not as of yet clear whether it was connected to the others.

The virus being spread is thought to be a version of the “Petya” ransomware and like the WannaCry virus that wreaked international havoc in May, it appears to take advantage of a Microsoft Windows flaw uncovered by the NSA and published online by hackers. This virus, however, seems to only be able to spread between directly connected networks, which is believed to be the reason the attack seemed to slow throughout the day Tuesday.

It’s still unclear as of now who is behind the attack.

Source: ABC

DNC hires ex-Uber engineer as its chief technology officer

The Democratic National Committee has selected ex-Twitter VP Raffi Krikorian as its chief technology officer. Krikorian was until recently the senior director of engineering at Uber’s Advanced Technologies Center, but left the company in February.

The DNC had plenty of cyber troubles during the election — most notably an email hack, information from which was subsequently posted by WikiLeaks. And boosting the DNC’s cyber defenses will clearly be one of Krikorian’s priorities. However, DNC Chair Tom Perez told Politico in January that he wanted an in-house cybersecurity officer to also work with state partners, not just the DNC. Pennsylvania Democrats were hit with ransomware in March and last year, the FBI suspected a number of Democrat officials’ cellphones had been hacked.

Some of Krikorian’s other roles might include some much-needed updates to the DNC’s use of technology in their campaigns. In the job posting for the CTO, the DNC made it clear they were looking for someone who could spur some innovation in how the party uses tech. The DNC hasn’t yet announced when Krikorian will start his new position.

Source: Recode

Artistic AI paints portraits of people who aren’t really there

Mike Tyka paints the portraits of people who don’t exist. The subjects of his ephemeral artwork are not born from any brush. Rather, they are sculpted — roughly — from the digital imagination of his computer’s neural network.

Faces are “interesting and we humans react strongly to them, we tend to read a lot into them,” Tyka explained over email. “I find I connect with them when I work with them, I’m curious about who they might be, if they existed.”

As such, Tyka has spent the past nine months or so developing the Portraits of Imaginary People project, which follows his earlier works, Inceptionism and The Groovik’s Cube. For Imaginary People, Tyka sought to use generative neural networks to create original portraits, much like the one Alexander Reben used to mimic Bob Ross’ speaking style.

To do so, Tyka turned to a machine learning technique known as a generative adversarial network (GAN). “I started experimenting with GANs for this installation I did with Refik Anadol, where we used the technique to generate imaginary historical documents from a large archive,” Tyka wrote. “After we finished that project (it opened in April in Istanbul), I started looking at faces again using the same techniques.”

If you want a generative model like a GAN to, say, draw you a picture of a cat, you’ll first have to get a huge data set of cat pictures and then train the model to create a picture of a cat with all the requisite features like ears, whiskers and a tail. In this case, Tyka utilized roughly 20,000 high-resolution portraits from Flickr as his base training data set.

That’s a good first step, but let’s say you want a realistic picture of a cat, not a digital doodle. To do that, you need to set up a second (adversarial) neural network, known as a discriminator, for your GAN. So while the first network (the generator) creates pictures of cats, the discriminator’s job is to compare those generated images against real-world samples (e.g., actual pictures of cats) and figure out if they’re fake or not. Based on each result, the system then goes back and tweaks the generator network’s parameters to make the output image appear more and more realistic.

If you’re only using a single, unconditioned GAN, the output image is typically only going to be in the 128×128 to 256×256 pixel range, Tyka explained. So to increase the size of these machine-generated images, he stacked multiple, separately trained GANs on top of one another. “The second stage is a superres GAN which [sic] is conditioned on the output of the former,” Tyka said. “I.e. in addition to the discriminator loss (which tries to make it look ‘real’) there is an additional term that makes sure the output is a plausible high-res version of the respective low-res input.” This second stage effectively increases the image resolution to 768×768 or 1024×1024 pixels.

By training the second-level (or even third-level) GAN on higher-resolution images of specific facial details like eye, hair and skin texture, it can act as an upscaler for the GANs stacked below it. Eventually Tyka wants to generate 4K-quality pictures, though he’s currently having difficulty finding a sufficiently robust data set for training such a system.

Getting the results you see here are easier said than done, however. There was plenty of work to complete before the first simulation ever ran. “GANs are hard to train and hard to control,” Tyka explained. “Grooming the input data is important, making sure all images are high-res, don’t have artifacts and are not drawings but real photos is time consuming.”

What’s more, keeping the adversarial networks in sync requires a fair amount of trial and error. “GANs are annoying because there isn’t a global objective function. The two networks are each other’s objective functions so to speak, so the goalposts are moving,” Tyka explained. “It’s hard to compare different runs with different parameters because there isn’t a good, stable metric for how well a particular net is doing.”

Still, Tyka’s desired end result for this project does not revolve around accuracy or fidelity. “The goal, like with many art projects, is to make compelling artwork which [sic] inspires or moves or makes you think,” he concluded. “That’s hard to quantify so I just follow my gut.”

Canada says court order to pull Google results applies worldwide

In 2012, Canadian manufacturer Equustek asked Google to remove search results relating to a court case against Datalink, a distributor of the former company’s network devices. While Google complied with the request, it only did so in Canada itself. The Supreme Court then ordered Google to remove search results pertaining to the issue in all countries Google operated in. Google appealed the decision, arguing that the order went against its own freedom of expression. The court has now rejected the company’s argument. The majority decision says that Canadian courts may in fact grant injunctions that compel compliance anywhere in the world.

“The internet has no borders – its natural habitat is global,” the Supreme Court wrote in its judgment. “The only way to ensure that the interlocutory injunction attained its objective was to have it apply where Google operates – globally.”

This isn’t the first time a country has told Google to comply globally. EU courts told Google to allow user requests to pull outdated or irrelevant information, while a court in France ruled that the company must extend the so-called “right to be forgotten” rule to global, not just regional, search results.

As reported by Reuters, the current Canadian ruling is opposed by civil liberties groups since it sets a precedent for internet censorship. “There is great risk that governments and commercial entities will see this ruling as justifying censorship requests that could result in perfectly legal and legitimate content disappearing off the web because of a court order in the opposite corner of the globe,” said David Christopher, a spokesperson for OpenMedia, a Canadian group that campaigns against censorship. Google cannot appeal the Supreme Court’s decision, but may apply for an alteration to the order when it has evidence that complying would violate other countries’ laws, according to Reuters.

Source: Reuters

Google Photos’ AI-powered sharing is now available

Google is making good on its promise of AI-assisted photo sharing. A Google Photos upgrade arriving this week uses machine learning to suggest pictures based on both your own sharing habits, the people in the photos, and whether or not they’re part of a “meaningful moment,” such as a party or a wedding. You might not have to remember to share photos of your best friend when you get home from a big weekend shindig. You can customize who receives the photos, of course, and fellow Google Photos users can get reminders to add their photos to the relevant album.

The update is also a big deal if you always want to share snapshots. There’s now an option to share your photo library, whether it’s the whole thing or snippets based on specific criteria. You can limit Photos to sharing pictures that include your partner, for example, or only those photos taken from a certain date onward.

Neither feature is flawless. What if your significant other has their back turned to the camera in an otherwise important shot? And while your friends won’t need Google Photos to receive suggested shares, the shared library clearly depends on everyone signing up. Still, this might be one of the more practical examples of how AI technology can help in everyday life. You won’t always have to remember to share photos when you get home — a machine will do much of the work for you.

Source: Google

Fox Sports brings its ‘virtual suite’ to Gold Cup VR broadcasts

Fox Sports is no stranger to VR. The network broadcasted Super Bowl 51 via a virtual suite of sorts, for instance. Now Fox is taking that further with the CONCACAF Gold Cup soccer championship. The footie action starts July 8th with the USA vs. Panama match at 4:30 PM Eastern, and will continue with two as-of-yet unannounced games thereafter. Basically, it’s combining that social space from American football with its previous work of broadcasting what everyone else around the world calls football, in VR.

The suite sounds a lot like what Fox had on offer for the Super Bowl earlier this year: multiple camera angles to choose from, there will be several 360-degree pre-show videos and replay features, too. The differences this time are that it’s sponsored by Buffalo Wild Wings and you can link up to watch with friends. So you should probably expect to see the corporate sports bar’s familiar black and yellow logo wherever you look.

If there’s a catch, it’s that you need a Facebook account to login and join your friends so you can watch from the social lounge together. No friends? No problem: the app will pair you with randoms too. Hanging out with strangers in VR was on your bucket list anyway, right?

Telegram will register with Russia but won’t share secure data

A few days ago, we reported that Russia’s communications regulator demanded that messaging app Telegram hand over information, including decrypted user messages, or risk being banned from the country. Now, Reuters reports that Telegram has agreed to register with the Russian government but will not hand over any user data or messages.

Russia claims it needs information from Telegram to put the app on a list of information distributors that officially operate within the country. While Pavel Durov, the founder of Telegram, says that the company is happy to register, he refuses to comply with any requests that violate Telegram’s confidentiality policy.

The reason behind Russia’s demands? It says that terrorists have been using the Telegram app to plan attacks, including a suicide bombing in April. While Telegram has been shutting down known terrorist channels on the network, it’s easy for them to create new ones.

This is just the latest conflict in the ongoing struggle between the need to protect users’ privacy and a desire to crack down on terrorist communication. For now, though, it appears as though users’ Telegram data is safe: Durov has been vehemently arguing in favor of privacy protection on Twitter and has made it clear the service won’t compromise on this issue.

Source: Reuters