Year: 2018

With Google’s I/O developer conference kicking off later today, Google is setting the scene for what it expects to be one of the big themes of the event: artificial intelligence. Today, the company rebranded the whole of its Google Research division as Google AI, with the old Google Research site now directing to a newly expanded Google AI site.

Google has over the years worked on a wide variety of other computing pursuits beyond AI, and all of that content will continue to exist within that new site, the company said. But the move signals how Google has increasingly focused a lot of its R&D on breaking new ground across the many facets of AI specifically, from technologies like computer vision, natural language processing, and neural networks, through to applications across virtually any and every business that Google currently and potentially touches, such as video, search and mobile apps, but also healthcare, automotive applications and other verticals.

That’s not just Google reflecting how the wider world of tech is evolving; it’s also a measure of how much Google has influenced it.

“From computer vision to healthcare research to AutoML, we have increasingly put emphasis on implementing machine learning techniques in nearly everything we do at Google,” writes Google’s Christian Howard in a post announcing the change. “Our research has been core to the development and integration of these systems into Google products and platforms. To better reflect this commitment, we’re unifying our efforts under “Google AI”, which encompasses all the state-of-the-art research happening across Google.”

Google seems to imply that further non-AI research will also be housed in the new AI hub, although it’s not really clear how much support and interest there will be for more work in other fields, or if everything now will need to have an AI angle to be progressed. (We’re asking and will update with a response when and if we get it.)

There is some interesting wider context for the rebranding. Just earlier today, Facebook also announced that one of its own research labs, its VR-focused Oculus Research, would be renamed the Facebook Reality Lab (notably no “virtual” in the name). Microsoft actually led the charge nearly two years ago when it recast its own Microsoft Research as Microsoft AI+Research — but it hasn’t (yet?) dropped the Research part of that name.

One big takeaway, I think, is that companies are trying to bring more of their own branding and more application focus into their quasi-academic research groups.

This could be happening here, too. Renaming Google Research into Google AI is a specific change that brings a division whose name might have previously had more general and academic connotations (“research”) into one that is now named much more specifically, and with an end product in mind (“AI”).

It’s a shift that could signify how the company is trying to reorient staff to move away from research for the sake of it, and towards more practical applications.

In the current race for talent and more business opportunities in the next generation of computing, Google is keeping its eye on the AI prize.

Town Hall Ventures launches today as the latest venture investment firm to take a stab at transforming the woefully broken U.S. health care market through technology.

What Town Hall has that many other firms trying to invest in healthcare technology don’t is a group of founders who have spent time in the trenches of the American healthcare system.

The firm is led by Andy Slavitt, the former Administrator of the Center for Medicaid and Medicare Service and the group executive vice president of Optum . Joining Slavitt are Trevor Price, the founder of Oxeon Holdings, an executive search and investment firm focused on life sciences, and David Whelan, the managing general partner of Oxeon Ventures — the predecessor to Town Hall.

Basically, it seems like Oxeon’s sort of spinning out its venture operations into a separate fund and has brought on Slavitt as a new top investor to manage things and add his years of knowledge working with the U.S. government’s healthcare system to the mix.

“Identifying talented entrepreneurs and helping them through the painstaking process of building a great company is something we are passionate about,” said Trevor Price, in a statement. “We believe we will see more entrepreneurs whose missions are to serve the tens of millions of people whose lives can be improved by innovations that allow them to receive the absolute best quality of care, whether being treated at home and/or in other comfortable and low-cost settings.”

While billions of venture dollars are heading towards the healthcare problem here in the U.S., the problems our system of services faces are so large — 120 million Americans spend around $1.2 trillion on health care — that there’s plenty of room for a number of competitors.

Joining the three founding partners are David Mishkin, a former director at Oxeon Partners, who will serve as the firm’s investment director; Natalie Davis, who worked with Slavitt as a senior advisor at CMS; and Lauren Robb, an investment associate who previously worked for Slavitt’s former company, Optum.

“As a nation we have significant health care infrastructure serving healthy populations while lower-income communities go underserved, leading to vastly poorer health outcomes,” said Slavitt. “We are at the beginning of a wave of innovation serving Medicare and Medicaid populations. Town Hall is being formed to help lead this massive and necessary shift. The answers are not always traditional; they involve investments in underlying systemic issues as well as innovative approaches that improve people’s health and well-being.”

The firm already has a few interesting investments under its belt — some through the portfolio created by Oxeon Partners.

According to a statement, portfolio companies include:

• Cityblock Health, Inc., a startup created with Alphabet’s Sidewalk Labs to provide primary care, behavioral health, and human services for cities (and an Oxeon portfolio company);
• Somatus, Inc., which is focused on creating new models of care for chronic kidney patients using technology;
• Welbe Health, LLC  a medical and social services provider for seniors who qualify for PACE;
• Aetion, Inc., an Oxeon portfolio company that sells real-world analytics and evidence to help biopharma companies and payers better understand how drugs work.

The pitch for founders is the staffing network that Town Hall can tap for young companies starting out to tackle huge problems in a massive market, according to executives who have worked with the firm.

“Without Town Hall and the resources and expertise they bring to bear, our goal of transforming care for urban and low-income populations would be moving far slower. In a short time, they have helped us roll out services in our first community, hire a first-class leadership team, and prepare ourselves for the daunting set of challenges that often substantially slow market entry in new health care organizations,” said Iyah Romm, the CEO of Cityblock Health, in a statement. “As a first-time entrepreneur, Town Hall surrounded me with a group of experienced leaders and operators across regulatory, management, fundraising, and technology unparalleled in health care,”

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 5 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 that 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.

Square, which has already made its way into retail stores and service-based businesses (think hair salons, massage therapists, etc), is officially getting into the restaurant business with the launch of Square for Restaurants. Square for Restaurants is a point-of-sale system that handles everything from menu updates, floor layouts, employee scheduling, performance tracking to tip splitting.

Usually, restaurants have “some old legacy thing or something else,” Square Seller Lead Alyssa Henry told me.

“Historically, we’ve not served this customer segment very well,” Henry said. “With Square for Restaurants, we’re excited to finally be able to serve this customer segment and deliver on a couple of key things that are core to Square but also highly valued by sellers of all types.”

This new product is designed to be fast, self-serve, elegant and cohesive, Henry said. It also integrates seamlessly into Square’s existing ecosystem that includes Payroll, Capital and more. Given Square’s ownership of on-demand food delivery startup Caviar, it’s no wonder why Square fully integrated Caviar into the point-of-sale system. This means restaurants don’t need to have separate tablet system for Caviar.

Square for Restaurant seating layout

“The omnichannel piece of restaurants is now truly coming into fruition with this integration,” Caviar Product Lead Gokul Rajaram told me. “We believe and we’ve seen restaurants increasingly becoming a multichannel or omnichannel platform where an increasing percent of their orders are coming from online channels — not just from diners coming into the store, but from delivery and pickup.”

Now, through the Square system, restaurants will be able to handle everything in one place. That means they can see their sales broken out by channel and understand what percentage of sales comes from delivery vs pickup vs in-restaurant dining, Rajaram said.

While these don’t yet exist, third-party applications from Postmates, UberEats and DoorDash could integrate into the Square POS. The issue right now for restaurants, according to Square, is the fact that delivering food for multiple services means a big of a tablet farm.

“The intention is that we can work with all of those because that’s what our sellers want,” Henry said.

While the product has been in beta, over 100 restaurants have used the product, including Bar Agricole in San Francisco and Greca in New York City. According to Square, the platform is relatively easy to set up. At launch, Square is charging $60 per month plus $40 per month for each additional POS set-up.

Microsoft and Red Hat are deepening their existing alliance around cloud computing. The two companies will now offer a managed version of OpenShift, Red Hat’s container application platform, on Microsoft Azure. This service will be jointly developed and managed by Microsoft and Red Hat and will be integrated into the overall Azure experience.

Red Hat OpenShift on Azure is meant to make it easier for enterprises to create hybrid container solutions that can span their on-premise networks and the cloud. That’ll give these companies the flexibility to move workloads around as needed and will give those companies that have bet on OpenShift the option to move their workloads close to the rest of Azure’s managed services like Cosmos DB or Microsoft’s suite of machine learning tools.

Microsoft’s Brendan Burns, one of the co-creators of Kubernetes, told me that the companies decided that this shouldn’t just be a service that runs on top of Azure and consumes the Azure APIs. Instead, the companies made the decision to build a native integration of OpenShift into Azure — and specifically the Azure Portal. “This is a first in class fully enterprise-supported application platform for containers,” he said. “This is going to be an experience where enterprises can have all the experience and support they expect.”

Red Hat VP for business development and architecture Mike Ferris echoed this and added that his company is seeing a lot of demand for managed services around containers.

Artificial intelligence has been a mainstay of the conversation in Silicon Valley these past few years, and now the technology is increasingly being discussed in policy circles in DC. Washington types see opportunities for AI to improve efficiency and increase economic growth, while at the same time, they have growing concerns around job automation and competitive threats from China and other countries.

Now, it appears the White House itself is getting involved in bringing together key American stakeholders to discuss AI and those opportunities and challenges. According to Tony Romm and Drew Harwell of the Washington Post, the White House intends to bring executives from major tech companies and other large corporations together on Thursday to discuss AI and how American companies can cooperate to take advantage of new advances in these technologies.

Among the confirmed guests are Facebook’s Jerome Pesenti, Amazon’s Rohit Prasad, and Intel’s CEO Brian Krzanich. While the event has many tech companies present, a total of 38 companies are expected to be in attendance including United Airlines and Ford.

AI policy has been top-of-mind for many policymakers around the world. French President Emmanuel Macron has announced a comprehensive national AI strategy, as has Canada, which has put together a research fund and a set of programs to attempt to build on the success of notable local AI researchers such as University of Toronto professor George Hinton, who is a major figure in deep learning.

But it is China that has increasingly drawn the attention and concern of U.S. policymakers. The country and its venture capitalists are outlaying billion of dollars to invest in the AI industry, and it has made leading in artificial intelligence one of the nation’s top priorities through its Made in China 2025 program and other reports. These plans are designed to coordinate various constituencies such as university researchers, scientists, companies, venture capitalists, and anyone else who might be able to assist in building out China’s AI capabilities.

In comparison, the United States has been remarkably uncoordinated when it comes to AI. While the government has released some strategic plans, it has mostly failed to follow through on coordinating more dollars toward artificial intelligence. As the New York Times noted in February, the White House has been remarkably silent on AI, despite the growing discussions around the technology.

That lack of engagement from policymakers has been fine — after all, the United States is the world leader in AI research. But with other nations pouring resources and talent into the space, DC policymakers are worried that the U.S. could suddenly find itself behind the frontier of research in the space, with particular repercussions for the defense industry.

Expect more news on this front in the coming months as DC’s various think tanks and analysts get their policy processes in motion.

Facebook has announced it has stopped accepting ads paid for by foreign entities that are related to a referendum vote in Ireland later this month, saying it’s acting to try to prevent outsiders from attempting to skew the vote. The referendum will decide whether to repeal or retain Ireland’s constitutional ban on abortion.

“Concerns have been raised about organisations and individuals based outside of Ireland trying to influence the outcome of the referendum on the Eighth Amendment to the Constitution of Ireland by buying ads on Facebook. This is an issue we have been thinking about for some time,” the company writes today on its Dublin blog.

“Today, as part of our efforts to help protect the integrity of elections and referendums from undue influence, we will begin rejecting ads related to the referendum if they are being run by advertisers based outside of Ireland.”

Facebook says it’s stopping foreign-funded ads because additional ad transparency and election integrity tools it has in the works — and is intending to roll out more widely, across its platform — will not be ready in time for Ireland’s Eighth Amendment vote, which will take place on May 25.

“What we are now doing for the referendum on the Eighth Amendment will allow us to operate as though these tools, which are not yet fully available, were in place today with respect to foreign referendum-related advertising. We feel the spirit of this approach is also consistent with the Irish electoral law that prohibits campaigns from accepting foreign donations,” Facebook writes.

“This change will apply to ads we determine to be coming from foreign entities which are attempting to influence the outcome of the vote on May 25. We do not intend to block campaigns and advocacy organisations in Ireland from using service providers outside of Ireland,” it adds.

The social media’s ad platform has been under increasing political scrutiny since revelations emerged about the extent of Kremlin-backed disinformation campaigns during the 2016 US presidential election. And last year Facebook admitted Kremlin-backed content — including, but not limited to, Facebook ads — may have reached as many as 126 million people during the election period.

Concerns have also been raised about the role of its platform during the UK’s 2016 referendum on EU membership — with an investigation into social media and campaign spending ongoing by the UK’s Electoral Commission, and another — by the UK’s data watchdog, the ICO — also looking more broadly at the use of data analytics for political purposes.

At the same time, a major Facebook data privacy scandal that erupted in March, after fresh details were published about the use of user data by a controversial political consultancy called Cambridge Analytica, has further dialed up the pressure on the company as lawmakers have turned their attention to the messy intersections of social media and politics.

Of course Facebook is by no means the only place online where all sorts of foreign agents have been caught seeking to influence opinions. But the Cambridge Analytica scandal has illustrated the powerful lure of the platform’s reach (and data holdings), as well as underlining how lax Facebook has historically been in controlling the messages people are paying it to target at its users.

In Ireland, the company had already fast-tracked the rollout of its ‘view ads’ ad transparency tool — ahead of a wider global rollout planned for this summer.

And last month policy staffers told a local parliamentary committee that the tool would help eliminate “foreign interference” in the upcoming referendum.

Although clearly Facebook has decided that an additional stop-gap measure — i.e. of rejecting foreign funded ads — was also needed given the timing (and indeed the sensitivity) — of the Eighth Amendment vote.

Last month Facebook also trailed plans to require advertisers that run popular Pages and/or are trying to run ads with political messages to verify their identity and location. But those advertiser verification steps do not appear to be ready in time for Ireland’s referendum. (Nor indeed were they in place for local elections in the UK earlier this month — although in a referendum the risks to democracy from a skewed vote are arguably higher, given there’s no established process for a re-vote in a few years’ time.)

The simpler-to-implement ‘view ads’ tool launched in Ireland on April 25, according to Facebook, which makes it the second market after Canada — where it began testing the feature.

The company claims the tool “enables Irish Facebook users to see all of the ads any advertiser is running on Facebook in Ireland at the same time” — though clearly ad visibility is not enough of a barrier against election fiddling on its own.

Facebook also says it will be using machine learning technology to help it identify ads that “should no longer be running”. And it’s supplementing these AI checks with human review, saying it’s built relationships with “political parties, groups representing both sides of the campaign and with the Transparent Referendum Initiative” — and is asking them to notify it if they have concerns about ad campaigns so it can assess and act on their reports, having established a dedicated reporting channel for this purpose.

Last month it also says it hosted an information session about its advertising and content policies for referendum campaign groups.

“We understand the sensitivity of this campaign and will be working hard to ensure neutrality at all stages. We are an open platform for people to express ideas and views on both sides of a debate. Our goal is simple: to help ensure a free, fair and transparent vote on this important issue,” it adds.

In addition to view ads and the decision to stop accepting foreign-funded referendum ads, Facebook says it is deploying its “Election Integrity Artificial Intelligence” for the vote in Ireland, as part of its efforts to identify fake accounts, misinformation and/or foreign interference — describing its approach as similar to what it did in advance of recent elections in France, Germany and Italy.

Last month its policy staffers also said it had set up an internal task force to handle the Ireland referendum.

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.

Uber sees a future where users can request a flying uber. And it’s what, as CEO Dara Khosrowshahi calls it, a big, bold bet. He says in an interview with CBS This Morning that big bold bets are what built Uber.

“We want to create the network around those vehicles so regular people can take these taxis in the air for longer distances when they want to avoid traffic at affordable prices,” said Khosrowshahi.

The goal is for the flying taxis to be driverless and hold four riders per vehicle to keep the cost lower for the passengers. Users would haul the air taxi from an app and then meet it at an uber rooftop facility. According to the CBS interview, Uber says the taxis will be relatively quiet thanks to multiple props and electric motors.

The company plans to launch trials as early as 2020.

Uber is set to unveil more details about its air taxi today and tomorrow at its Uber Elevate conference.