Between Microsoft Build and Google I/O, there are probably more people saying “AI” this week than any previous week in history. But the AI those companies deploy tends to live off in a cloud somewhere — XNOR puts it on devices that may not even be capable of an internet connection. The startup has just pulled in $12 million to continue its pursuit of bringing AI to the edge.

I wrote about the company when it spun off of Seattle-based, Paul Allen-backed AI2; its product is essentially a proprietary method of rendering machine learning models in terms of operations that can be performed quickly by nearly any processor. The speed, memory and power savings are huge, enabling devices with bargain-bin CPUs to perform serious tasks like real-time object recognition and tracking that normally take serious processing chops to achieve.

Since its debut it took $2.6 million in seed funding and has now filled up its A round, led by Madrona Venture Group, along with NGP Capital, Autotech Ventures and Catapult Ventures.

“AI has done great,” co-founder Ali Farhadi told me, “but for it to become revolutionary it needs to scale beyond where it is right now.”

The fundamental problem, he said, is that AI is too expensive — both in terms of processing time and in money required.

Nearly all major “AI” products do their magic by means of huge banks of computers in the cloud. You send your image or voice snippet or whatever, it does the processing with a machine learning model hosted in some data center, then sends the results back.

For a lot of stuff, that’s fine. It’s okay if Alexa responds in a second or two, or if your images get enhanced with metadata over a period of hours while you’re not paying attention. But if you need a result not just in a second, but in a hundredth of a second, there’s no time for the cloud. And increasingly, there’s no need.

XNOR’s technique allows things like computer vision and voice recognition to be stored and run on devices with extremely limited processing power and RAM. And we’re talking Raspberry Pi Zero here, not just like an older iPhone.

If you wanted to have a camera or smart home type device in every room of your home, monitoring for voices, responding to commands, sending its video feed in to watch for unauthorized visitors or emergency situations — that constant pipe to the cloud starts getting crowded real fast. Better not to send it at all.

This has the pleasant byproduct of not requiring what might be personal data to some cloud server, where you have to trust that it won’t be stored or used against your will. If the data is processed entirely on the device, it’s never shared with third parties. That’s an increasingly attractive proposition.

Developing a model for edge computing isn’t cheap, though. Although AI developers are multiplying, comparatively few are trying to run on resource-limited devices like old phones or cheap security cameras.

XNOR’s model lets a developer or manufacturer plug in a few basic attributes and get a model pre-trained for their needs.

Say you’re the cheap security camera maker; you need to recognize people and pets and fires, but not cars or boats or plants, you’re using such and such ARM core and camera and you need to render at five frames per second but only have 128 MB of RAM to work with. Ding — here’s your model.

Or say you’re a parking lot company and you need to recognize empty spots, license plates and people lurking suspiciously. You’ve got such and such a setup. Ding — here’s your model.

These AI agents can be dropped into various code bases fairly easily and never need to phone home or have their data audited or updated, they’ll just run like greased lightning on the platform. Farhadi told me they’ve established the most common use cases and devices through research and feedback, and many customers should be able to grab an “off the shelf” model just like that. That’s Phase 1, as he called it, and should be launching this fall.

Phase 2 (in early 2019) will allow for more customization, so for example if your parking lot model becomes a police parking lot model and needs to recognize a specific set of cars and people, or you’re using proprietary hardware not on the list. New models will be able to be trained up on demand.

And Phase 3 is taking models that normally run on cloud infrastructure and adapting and “XNORifying” them for edge deployment. No timeline on that one.

Although the technology lends itself in some ways to the needs of self-driving cars, Farhadi told me they aren’t going after that sector — yet. It’s still essentially in the prototype phase, he said, and creators of autonomous vehicles are currently trying to prove the idea works fundamentally, not trying to optimize and deliver it at lower cost.

Edge-based AI models will surely be increasingly important as the efficiency of algorithms improves, the power of devices rises and the demand for quick-turnaround applications grows. XNOR seems to be among the vanguard in this emerging area of the field, but you can almost certainly expect competition to expand along with the market.

True friends share everything, from purses to lawnmowers to the limitless energy they pull from the sun, wind, and deep underground. Green Power Exchange aims to make one of those sharing scenarios a bit simpler.

Green Power Exchange or GPX is running an ICO to create a coin that lets you trade energy. While there are a number of efforts on this front, GPX is fairly far along and already has 62 solar sites signed up to try the tech when it launches. Founder Christian Wentzel sees the project as a solution to the energy monopolies.

“The utility and power company either is a direct or quasi-monopoly, leaving no customer choice,” he said. “We want to change that and save consumers a ton of money while improving the environment.”

“Solar and wind power is cheap, so we want people to use more of it.”

The system essentially greases the wheels of energy trading. Whole current solutions require complex transactions and contracts, this solution lets energy generators sell their excess energy to other people with a minimum of fuss. The team plans to launch in 2019. The have created two coins, the GPX and the GET. The GPX is a liquid coin that can be used to buy and sell power while the GET is tied to the actual generation of energy. You would buy and sell GPX, for example, and manage your energy purchases using the GET token.

GPX’s goals are noble and these sorts of markets are definitely headed toward the blockchain. ” One of the largest barriers to renewable power development in these markets is the complexity of PPA negotiations and the rigidity of the agreements,” said Wentzel. It’s at least clear that something like GPX could help.

As we plunge into our baffling future, it is believed that, at some point, we will be trading in cryptographically secure kittens, monsters, and playing cards. While it is unclear why this will happen, Rare Bits and their new service, Fan Bits, is ready for the oncoming rush.

Co-founded by Dave Pekar, Amitt Mahajan and Danny Lee (who met after selling their gaming startups to Zynga) and Payom Dousti (formerly of fintech VC fund 1/0 Capital), the company trades in digital goods and has built a blockchain-based solution for buying and selling digital collectables. Lee brought in a team of ex-Zynga and other digital platform creators to build a blockchain-based solution for buying and selling digital collectables. For example, on Rare Bits you can buy this monster and battle it against other monsters on the blockchain. Further, with their new platform called Fan Bits, you can buy actual collectables that are tied to the blockchain. For example, you can sell collectible cards and give some of the proceeds to charity. If the new owner resells those cards then some of the resell price also goes to charity, an interesting if slightly intrusive use of smart contracts.

The team has raised $6 million in Series A. Fan Bits launches on May 17.

“To date, collectible content has only been created by developers for their own dapps – which I suppose could be considered our competition,” said Lee. “Fan Bits is the first to let anyone, especially people who are not technical, to create collectibles. It will create an abundance of supply that didn’t exist before.”

“We started Rare Bits to let people buy, sell, and discover crypto assets. We believe that assets on the blockchain mark a fundamental shift in how we own and exchange property. Our overall mission is to enable the worldwide exchange of online and offline property on the blockchain,” he said.

Lee sees this as a Trojan horse of sorts, allowing non tech-savvy creators sell digital art and designs online without having to understand the vagaries of blockchain.

“For creators, it’s a DIY platform to turn their content into unique collectibles and earn Ethereum on every sale,” he said. “For the first time, a creator can go from idea to published cryptocollectible on a live marketplace without having to have any technical knowledge.”

Given the popularity of other digital collectables – including in-game gear for many multi-player games – things look like they’re going to get pretty interesting in the next few years.

There are times where voice-driven systems don’t work all that well because of background noise or other voices. That’s because it’s hard for machines (and humans) to pull out a particular voice when there are many others speaking. This is sometimes called ‘the cocktail party problem.’ Yobe was created out of research at MIT on how to solve this issue, and today it announced $1.8 million in seed funding.

The investment comes from Clique Capital Partners, a $100 million fund created specifically to fund innovative voice technology. Yobe had previously received $790,000 in the form of a National Science Foundation SBIR grant in 2016.

Company co-founder and CEO Ken Sutton says Yobe is solving an entrenched problem identifying a particular voice in noise. That means for instance if you are at a party and you want Alexa to play a Spotify playlist, you could (in theory at least), say the wake word from across a crowded room, give the playlist command and the device would execute it in spite of the noise. That’s because Yobe can pinpoint a voice based on biometric markers, aggressively enhance the volume and then use AI to smooth it out.

As Sutton says, most of these voice interface technologies fail in this situation because they can’t distinguish your voice from the background noise, but Yobe is supposed to solve this.

Sutton made clear the research phase is done and the funding is about getting ready to go to market. “The capital raised is not to continue R&D. The capital raised is to streamline and optimize [the technology] for deployment. We will be in market with a product to sell in 30 days, and all of the usual suspects are lined up and waiting for us to call with a live demo,” Sutton told TechCrunch.

Ultimately the company hopes to license its technology to chip or phone manufacturers and others in a scenario not unlike Dolby. Sutton acknowledges there are many use cases for a  solution that could identify a specific voice in noise or among other voices such as law enforcement, hearing aid manufacturers and meeting transcription services. You could even use voice as a biometric marker for authentication purposes. But the company has decided to focus its early efforts on voice-driven devices as an initial go-to market strategy.

The company was founded by Sutton and Dr. S. Hamid Nawab, an MIT PhD and researcher, whose work has focused on applying AI to signal processing. Nawab is the company’s Chief Scientist.

You can watch this video demo to see how Yobe works:

According to Transparency Market Research, the global sales performance management market grew 17.1% annually from $2.3 billion in 2017 to $8.1 billion by the end of 2025. Utah-based InsideSales.com...

___

Dreamlines, which claims to be Europe’s largest online travel agency specialising in cruise-related travel, is disclosing that it has raised €45 million in Series E funding. The round is led by Princeville Global, with participation from existing investors that include Holtzbrinck Ventures, Target Global, Dimaventures, Hasso Plattner Ventures, TruVenturo, and Rocket Internet’s Global Founders Capital.

Founded by Felix Schneider in 2012, Hamburg-based Dreamlines can be thought of as a Booking.com or Expedia but for cruise holidays and other cruise line type travel. The company connects customers to what it says is the largest portfolio of cruises around the world, including holiday packages exclusive to Dreamlines.

Meanwhile, unlike other forms of holiday and travel, the cruise industry is only more recently being digitised, a sentiment echoed by Emmanuel DeSousa, Managing Partner of Princeville Global, who joins the Dreamlines board.

“The cruise industry is the last sizable, global travel segment to be disrupted by a tech-focused online booking platform,” he says. “Under the leadership of its visionary founders, Dreamlines is uniquely positioned to continue transforming the cruise industry to an online model, leading in Europe and expanding around the world”.

To that end, Dreamlines says the investment will support its continued growth and international expansion. The company currently operates in 10 countries, partnering with over 100 cruise operators, and has raised around €110 million to date.

Adds Christian Saller, General Partner at HV Holtzbrinck Ventures and the Dreamlines chairman: “As an early investor, HV Holtzbrinck Ventures has seen Dreamlines grow by a factor ten since its initial investment into the European market leader. The new investment will allow Dreamlines to continue this success story”.

Sramana Mitra: It sounds like when you did the cloud strategy, you almost did it with a different product. Even though the technology was the same, the positioning was quite different. The board...

___

Going to the dentist can be anxiety-inducing. Unfortunately, it was no different for me last week when I went to discuss Uniform Teeth’s recent $4 million seed funding round from Lerer Hippeau, Refactor Capital, Founder’s Fund and Slow Ventures.

Uniform Teeth is a clear teeth aligner startup that competes with the likes of Invisalign and Smile Direct Club. The startup takes a One Medical-like approach in that it provides real, licensed orthodontists to see you and treat your bite.

“For us, we’re really focused on transforming the orthodontic experience,” Uniform Teeth CEO Meghan Jewitt told me at the startup’s flagship dental office in San Francisco. “There are a lot of health care companies out there that are taking areas that aren’t very customer-centric.”

Jewitt, who spent a couple of years at One Medical as director of operations, pointed to One Medical, Oscar Insurance and 23andMe as examples of companies taking a very customer-centric approach.

“We are really interested in doing the same for the orthodontics space,” she said.

Ahead of the first visit, patients use the Uniform app to take photos of their teeth and their bite. During the initial visit, patients receive a panoramic scan and 3D imaging to confirm what type of work needs to be done.

Last week during my visit, Jewitt and Uniform Teeth co-founder Dr. Kjeld Aamodt showed me the technology Uniform uses for its patient evaluations.

In the GIF above, you can see I received a 3D panoramic X-ray. The process took about 10 seconds and it’s about the same exposure to X-rays as a flight from San Francisco to Los Angeles, Dr. Aamodt said.

“With that information, we’re able to see the health of your roots, your teeth, the bone, your jaw joints, check for anything that could get worse during treatment,” Dr. Aamodt said.

Below, you can see the 3D scan.

Next is looking in-between the teeth. From here, the idea is to get a much more holistic view, Dr. Aamodt said. This is where things got interesting.

If you look at the bottom left of the photo, under my back bottom tooth, you can see a dark circle below the tooth. Dr. Aamodt gently pointed that out to me.

“That tells me there’s bacteria living inside of your jaw,” he explained. “A lot of people have this. It’s pretty common so don’t beat yourself up for it.”

This is when he told me I’d likely need to get a root canal to get rid of it. Mild panic ensued.

Dr. Aamodt went on to explain that, if I were a patient of his looking to get my teeth straightened, he would recommend that I first get the root canal before any teeth movement. That’s because, he explained, moving teeth at that point could potentially result in further infection.

“The concern about that is when we move a tooth with that, the infection will get worse and you could risk losing that tooth,” he told me.

Although I was freaking out internally, I continued to move ahead with the process. Next up was the 3D scan, which results in something fancy called a 3D Cone Beam Computed Tomography. This, Dr. Aamodt said, is what really sets Uniform Teeth apart in precision tooth movement.

This process takes the place of dental impressions, which are made by biting into a tray with gooey material. I didn’t feel like getting my bottom teeth scanned, but below is what the top looks like.

At this point Uniform Teeth would share its recommendations with the patient. My personal recommendation was to go see my dentist and, if I’m interested in straightening my teeth, come back once my roots are in a healthy enough state.

From there, I’d receive a custom treatment plan that combines the X-ray plus 3D scan to predict how my teeth will move. After receiving the clear aligners in a couple of weeks, I’d check in with Dr. Aamodt every week via the mobile app. If something were to come up, I could always set up an in-person appointment. Most people average about two to three visits in total, Jewitt said. All of that would add up to about $3,500.

The reason Uniform Teeth requires in-office visits is because 75 percent or more of the cases require additional procedures. For example, some people require small, tooth-colored attachments to be placed onto the clear aligners. Those attachments can help move teeth in a more advanced way, Dr. Aamodt said.

“If you don’t have these, then you can tip some teeth but you can’t do all of the things to help improve the bite, to create a really lasting, beautiful, healthy smile,” he explained.

Uniform Teeth currently treats patients in San Francisco, but intends to open additional offices nationwide next year. As the company expands, the plan is to bring on board more full-time orthodontists.

“Right now, we’re an employment-based model and we’d really like to continue that because it allows us to control the experience and deliver a really high-quality service,” Jewitt said.

A lot of companies say they care about the customer when, in reality, they just care about making money. But I genuinely believe Uniform Teeth does care. After I left with my tail between my legs that day, I called my dentist to set up an appointment. The following day, my dentist confirmed what Dr. Aamodt found and proceeded to set me up to get a root canal. A few days later, Dr. Aamodt checked in with me via the mobile app to ask me how I was doing and to make sure I was getting it treated. I was pleased to let him know, as Olivia Pope likes to say, “It’s handled.”

Dave Hornik: In the venture-backed world, there have been very few apps that have turned into big interesting things. There are big games that have gotten very large. Those are tricky businesses...

___

Printify, a startup all the way from Riga, Latvia, has raised $1 million in seed funding led by Google AdSense pioneer Gokul Rajaram as it looks to expand its services in the U.S. and build out its team in Latvia.

Today, roughly 50,000 e-commerce stores use Printify’s services for printing-on-demand, according to the company.

Together with Lumi, which can handle packaging for consumer-facing startups, Printify is making the notion of becoming a brand as seamless as possible by taking much of a vendor’s legwork out of the equation.

The company keeps its customers’ billing information on file and links with the back-end ordering system of almost any e-commerce platform.

When an order comes in, Printify gets an API notification to begin working on a product. The company then sends print-ready files to an on-demand manufacturer that can print a design and ship a product within 24 hours.

Printify founder James Berdigans came up with the idea for the company after starting a business making accessories for Apple products.

“We wanted to print custom phone cases and we thought we’d find an on-demand manufacturer and it would be easy,” says Berdigans. That’s when Printify was born.

The company first integrated through Shopify and began to see some traction in November 2015, but because the company didn’t own its own manufacturing, quality became a concern.

In 2016, Berdigans pivoted to the marketplace model because he saw demand coming from a growing number of small business owners like himself that needed access to a selection of quality printing houses.

New direct-to-garment printing and digital printing on t-shirts is going to grow very rapidly over the next few years, according to Berdigans.

A graduate from the 500 Startups accelerator program, the company is already profitable and hoping to capitalize on its success with this new funding to expand its engineering team.

“In just two years, Printify has become profitable and is growing at a rapid pace. With this
investment, we plan to double our Riga team in 2018. That will create at least 30 new jobs,
most of which will be in programming, design, and customer support,” said Printify co-founder
Artis Kehris in a statement.

Purple Carrot announced this morning that it has raised $4 million in strategic funding from Fresh Del Monte Produce.

The company, which delivers completely plant-based (vegan) meal kits to subscribers, was founded in 2014. It has, in part, gotten attention through celebrity involvement, first by enlisting food writer Mark Bittman as its chief innovation officer (Bittman departed in 2016), then by partnering with football star and notorious strawberry hater Tom Brady to launch TB12 meal kits.

Purple Carrot had previously raised $6 million in funding, according to Crunchbase. The company says that this new investment will allow it to improve its supply chain, get access to more retail opportunities and explore expansion into other categories.

“Securing this strategic investment from Fresh Del Monte is a huge validation of our business model, and an important step forward for our company,” said Purple Carrot founder and CEO Andy Levitt in the announcement. “Helping people eat more plant-based foods represents our differentiated, purpose-driven commitment to making the planet and the people who live on it healthier.”

Fresh Del Monte (which is the company behind Del Monte pineapples and other produce) is just the latest established player in the food and grocery world to invest in meal delivery startups. Last year, for example, Unilever backed Sun Basket and Nestlé led a $77 million round for Freshly.

Entrepreneurs are invited to the 399th FREE online 1Mby1M mentoring roundtable on Thursday, May 10, 2018, at 8 a.m. PDT/11 a.m. EDT/8:30 p.m. India IST. If you are a serious entrepreneur, register to...

___

Sramana Mitra: What you’re saying is that you are open to doing even pre-seed investments? Greg Sands: Yes. They tend to be in more experienced founders or people that we’ve known for a longer period...

___

According to a recent market report, the global visual analytics market is expected to grow at a CAGR of 22% to reach $7.7 billion in 2023. Data analytics firm Tableau (NYSE: DATA) is a leading...

___

Machine learning may be the tool de jour for everything from particle physics to recreating the human voice, but it’s not exactly the easiest field to get into. Despite the complexities of video editing and sound design, we have UIs that let even a curious kid dabble in them — so why not with machine learning? That’s the goal of Lobe, a startup and platform that genuinely seems to have made AI models as simple to put together as LEGO bricks.

I talked with Mike Matas, one of Lobe’s co-founders and the designer behind many a popular digital interface, about the platform and his motivations for creating it.

“There’s been a lot of situations where people have kind of thought about AI and have these cool ideas, but they can’t execute them,” he said. “So those ideas just get like shed, unless you have access to an AI team.”

This happened to him, too, he explained.

“I started researching because I wanted to see if I could use it myself. And there’s this hard to break through veneer of words and frameworks and mathematics — but once you get through that, the concepts are actually really intuitive. In fact even more intuitive than regular programming, because you’re teaching the machine like you teach a person.”

But like the hard shell of jargon, existing tools were also rough on the edges — powerful and functional, but much more like learning a development environment than playing around in Photoshop or Logic.

“You need to know how to piece these things together, there are lots of things you need to download. I’m one of those people who if I have to do a lot of work, download a bunch of frameworks, I just give up,” he said. “So as a UI designer I saw the opportunity to take something that’s really complicated and reframe it in a way that’s understandable.”

Lobe, which Matas created with his co-founders Markus Beissinger and Adam Menges, takes the concepts of machine learning, things like feature extraction and labeling, and puts them in a simple, intuitive visual interface.

As demonstrated in a video tour of the platform, you can make an app that recognizes hand gestures and matches them to emoji without ever seeing a line of code, let alone writing one. All the relevant information is there, and you can drill down to the nitty gritty if you want, but you don’t have to. The ease and speed with which new applications can be designed and experimented with could open up the field to people who see the potential of the tools but lack the technical know-how.

He compared the situation to the early days of PCs, when computer scientists and engineers were the only ones who knew how to operate them. “They were the only people able to use them, so they were they only people able to come up with ideas about how to use them,” he said. But by the late ’80s, computers had been transformed into creative tools, largely because of improvements to the UI.

Matas expects a similar flood of applications, even beyond the many we’ve already seen, as the barrier to entry drops.

“People outside the data science community are going to think about how to apply this to their field,” he said, and unlike before, they’ll be able to create a working model themselves.

A raft of examples on the site show how a few simple modules can give rise to all kinds of interesting applications: reading lips, tracking objects’ positions and angles, understanding gestures, generating realistic flower petals. Petals? Well, why not? You need data to feed the system, of course, but doing something novel with it is no longer the hard part.

And in keeping with the machine learning community’s commitment to openness and sharing, Lobe models aren’t some proprietary thing you can only operate on the site or via the API. “Architecturally we’re built on top of open standards like Tensorflow,” Matas said. Do the training on Lobe, test it and tweak it on Lobe, then compile it down to whatever platform you want and take it to go.

Right now the site is in closed beta. “We’ve been overwhelmed with responses, so clearly it’s resonating with people,” Matas said. “We’re going to slowly let people in, it’s going to start pretty small. I hope we’re not getting ahead of ourselves.”

Shenzhen-based home robotics company UBTECH announced this week that it has closed a massive $820 million Series C. The round, led by Tencent and a whole slew of other investors, follows a $100 million Series B and $20 million Series C.

The bipedal robotics maker certainly isn’t a household name here in the States — though the company’s taken a few baby steps over here, including a walking Stormtrooper robot released alongside The Last Jedi. It also debuted a “robotic butler” with a tablet face back at CES in January that seemed more proof of concept than shipping product — though UBTECH has promised a broad 2019 release date.

CEO James Zhou has promised the company will use this huge backing to accelerate its vision of bringing robots into the home.

“As technology evolves to include more voice and touch capabilities, people need new devices that communicate and interact more naturally and intuitively at home, at school and at work,” he said in a release tied to the announcement. “While trends in robots and robotics are developing, no company has yet stepped forward with the resources, vision and products ecosystem to transform robot fantasy and fiction into robot reality. UBTECH is bringing this reality to life by expanding the possibilities for innovation.”

The company says the funding will go into R&D, hires and expanding its global footprint. UBTECH is one of a number of companies pushing home robotics beyond the Roomba, including a rumored upcoming device from Amazon. The technology has been a tough nut to crack in the preceding decades, but an $820 million investment certainly couldn’t hurt.