Month: June 2019

MIT’s Computer Science and Artificial Intelligence Lab has developed a new deep learning-based AI prediction model that can anticipate the development of breast cancer up to five years in advance. Researchers working on the product also recognized that other similar projects have often had inherent bias because they were based overwhelmingly on white patient populations, and specifically designed their own model so that it is informed by “more equitable” data that ensures it’s “equally accurate for white and black women.”

That’s key, MIT notes in a blog post, because black women are more than 42 percent more likely than white women to die from breast cancer, and one contributing factor could be that they aren’t as well-served by current early detection techniques. MIT says that its work in developing this technique was aimed specifically at making the assessment of health risks of this nature more accurate for minorities, who are often not well represented in development of deep learning models. The issue of algorithmic bias is a focus of a lot of industry research and even newer products forthcoming from technology companies working on deploying AI in the field.

This MIT tool, which is trained on mammograms and patient outcomes (eventual development of cancer being the key one) from over 60,000 patients (with over 90,000 mammograms total) from the Massachusetts General Hospital, starts from the data and uses deep learning to identify patters that would not be apparent or even observable by human clinicians. Because it’s not based on existing assumptions or received knowledge about risk factors, which are at best a suggestive framework, the results have so far shown to be far more accurate, especially at predictive, pre-diagnosis discovery.

Overall, the project is intended to help healthcare professionals put together the right screening program for individuals in their care and eliminate the heartbreaking and all-too common outcome of late diagnosis. MIT hopes the technique can also be used to improve detection of other diseases that have similar problems with existing risk models with far too many gaps and lower degrees of accuracy.

Over the past year, Songtrust has added another 55,000 artists to its rights management service.

The company, a subsidiary of Downtown Music Publishing, a publishing and rights management firm that manages rights for artists such as John Lennon, One Direction and Santigold, now has 205,000 artists on its roster and has 2 million songs it tracks.

The company has also opened three offices in Atlanta, Los Angeles, and Nashville to complement existing locations in New York, London and Amsterdam.

The company’s growth follows that of a music industry that continues to enjoy a renaissance (at least in terms of dollars spent).

The global recorded music market grew 9.7% in 2018 to $19.1 billion, according to data from the International Federation of the Phonographic Industry (which has been tracking the industry since the days when the dominant technology was the record player).

Much of that growth is now coming from streaming, the IFPI reports, with streaming revenues growing 34% year over year and accounting for 47% of total revenue thanks to paid subscription services. There were 255 million users of paid services by the end of 2018 — and Songtrust can attribute much of its growth to the opacity in how that money makes its way back to artists.

Increasingly, those artists are having to track their performance in international markets as well. Latin America continues to be the fastest gorwing region for music consumption, followed by Asia and Australasia. Most of that growth is due to K-Pop, since South Korea accounts for 17.9% growth in money spent alone.

All of this movement shows no sign of abating, according to the bankers that track these kinds of things. Goldman Sachs recently projected that the industry could grow to over $130 billion in revenue over the next decade.

Meme creators have never gotten their fair share. Remixed and reshared across the web, their jokes props up social networks like Instagram and Twitter that pay back none of their ad revenue to artists and comedians. But 300 million monthly user meme and storytelling app Imgur wants to pioneer a way to pay creators per second that people view their content.

Today Imgur announces that it’s raised a $20 million venture equity round from Coil, a micropayment tool for creators that Imgur has agreed to build into its service. Imgur will eventually launch a premium membership with exclusive features and content reserved for Coil subscribers. Users pay Coil a fixed monthly fee, install its browser extension, the Ripple XRP cryptocurrency is used to route assets around, and then Coil pays creators per second that the subscriber spends consuming their content at a rate of 36 cents per hour. Imgur and Coil will earn a cut too, diversifying the meme network’s revenue beyond ads.

“Imgur began in 2009 as a gift to the internet. Over the last 10 years we’ve built one of the largest, most positive online communities, based on our core value to ‘give more than we take’” says Alan Schaaf, founder and CEO of Imgur. The startup bootstrapped for its first five years before raising a $40 million Series A from Andreessen Horowitz and Reddit. It’s grown into the premier place to browse ‘meme dumps’ of 50+ funny images and GIFs, as well as art, science, and inspirational tales.

While the new round brings in fewer dollars, Schaaf explains that Imgur raised at a valuation that’s “higher than last time. Our investors are happy with the valuation. This is a really exciting strategic partnership.” Coil founder and CEO Stefan Thomas who was formerly the CTO of cryptocurrency company Ripple Labs will join Imgur’s board. Coil received the money it’s investing in Imgur from Ripple Labs’ Xpring Initiative, which aims to fund proliferation of the Ripple XRP ecosystem, though Imgur received US dollars in the funding deal.

Thomas tells me that “There’s no built in business model” as part of the web. Publishers and platforms “either make money with ads or with subscriptions. The problem is that only works when you have huge scale” that can bring along societal problems as we’ve seen with Facebook. Coil will “hopefully offer a third potential business model for the internet and offer a way for creators to get paid.”

Founded last year, Coil’s $5 per month subscription is now in open beta, and it provides extensions for Chrome and Firefox as it tries to get baked into browsers natively. Unlike Patreon where you pick a few creators and choose how much to pay each every month, Coil lets you browse content from as many creators as you want and it pays them appropriately. Sites like Imgur can code in tags to their pages that tell Coil’s Web Monetization API who to send money to.

The challenge for Imgur will be avoiding the cannibalization of its existing content to the detriment of its non-paying users who’ve always known it to be free. “We’re in the business of making the internet better. We do not plan on taking anything away for the community” Schaaf insists. That means it will have to recruit new creators and add bonus features that are reserved for Coil subscribers without making the rest of its 300 million users feel deprived.

It’s surprising thT meme culture hasn’t spawned more dedicated apps. Decade-old Imgur precedes the explosion in popularity of bite-sized internet content. But rather than just host memes like Instagram, Imgur has built its own meme creation tools. If Imgur and Coil can prove users are willing to pay for quick hits of entertainment and creators can be fairly compensated, they could inspire more apps to help content makers turn their passion into a profession…or at least a nice side hustle.

Apple today is introducing a new section in its Apple News application for iOS, iPad and Mac that’s designed to familiarize voters with the 20 U.S. Democratic presidential candidates ahead of the first Democratic debates hosted on June 26 and June 27 by NBC News, MSNBC, and Telemundo in Miami, Florida. The new guide is meant to provide a single place where readers can learn about a candidate’s biography, experience, and current position on key issues, among other things.

It will also feature photos and videos, along with recent coverage from trusted news sources.

Apple says that it will leverage a diverse set of news sources to provide this information, including ABC News, Axios, CNN, Fox News, NBC News, Politico, The Hill, The New York Times, The Wall Street Journal, The Washington Post, TIME, USA Today, Vox and others.

The candidate information is curated and organized by Apple’s team of News editors, and will be found in the Top Stories section of its Apple News app.

As the debates begin, the section will expand from being only a guide to candidates to being a hub for updates from the debates, too. It will then include articles and video highlights from NBC News, as well as fact-checking, reactions, and key onstage moments and takeaways, says Apple in an announcement about the new feature.

The hub will continue to be updated after the debates, as well, with more news throughout the campaign.

In addition, users can personalize their experience by tracking their favorite candidates in Apple News. To do so, they’ll just click the “follow” button for the candidate in the app, then will receive any breaking news around that candidate as well as see ongoing news coverage about the candidate appear in their Apple News Today feed and elsewhere within the News app.

This isn’t the first time Apple has involved itself in helping curate and organize election-related information. Most recently, it launched a real-time news hub for the 2018 Midterms.

The company has taken the opposite approach from Facebook when it comes to providing easy access to news and information to its users. Facebook fired its news editors then turned its “trends” section over to an algorithm, before giving up and killing it altogether. But Apple has instead hired an editorial team of former journalists to organize and curate the news that tens of millions of people read. The News team doesn’t write its own stories, but plays a large role in selecting the stories that people first see when they open the app.

The result is an easy-to-use app designed with Apple’s aesthetic, where journalism gets top billing — not clickbait, viral news, and intentional sources of disinformation.

“The 2020 Democratic field is complex, and we want to offer Apple News readers a trusted place to learn more about candidates they’re familiar with and those they may be hearing about for the first time,” said Lauren Kern, editor-in-chief of Apple News. “The candidate guide in Apple News is a robust and reliable resource, connecting readers to valuable at-a-glance information and to great journalism from our partners.”

Newly released documents reveal the National Security Agency improperly collected Americans’ call records for a second time, just months after the agency was forced to purge hundreds of millions of collected calls and text records it unlawfully obtained.

The document, obtained by the American Civil Liberties Union, shows the NSA had collected a “larger than expected” number of call detail records from one of the U.S. phone providers, though the redacted document did not reveal which provider nor how many records were improperly collected.

The document said the erroneously collected call detail records were “not authorized” by the orders issued by the Foreign Intelligence Surveillance Court, which authorizes and oversees the U.S. government’s surveillance activities.

Greg Julian, a spokesperson for the NSA, confirmed the report in an email to TechCrunch, saying the agency “identified additional data integrity and compliance concerns caused by the unique complexities of using company-generated business records for intelligence purposes.”

NSA said the issues were “addressed and reported” to the agency’s overseers, but did not comment further on the violations as they involve operational matters.

The ACLU called on lawmakers to investigate the improper collection and to shut down the program altogether.

“These documents further confirm that this surveillance program is beyond redemption and a privacy and civil liberties disaster,” said Patrick Toomey, a staff attorney with the ACLU’s National Security Project. “The NSA’s collection of Americans’ call records is too sweeping, the compliance problems too many, and evidence of the program’s value all but nonexistent.”

“There is no justification for leaving this surveillance power in the NSA’s hands,” he said.

Under the government’s so-called Section 215 powers, the NSA collects millions of phone records every year by compelling U.S. phone giants to turn over daily records, a classified program first revealed in a secret court order compelling Verizon — which owns TechCrunch — from documents leaked by whistleblower Edward Snowden. Those call records include the phone numbers of those communicating and when — though not the contents — which the agency uses to make connections between targets of interest.

But the government was forced to curtail the phone records collection program in 2015 following the introduction of the Freedom Act, the only law passed by Congress since the Snowden revelations which successfully reined in what critics said was the NSA’s vast surveillance powers.

In recent years, the number of call records has gone down but not gone away completely. In its last transparency report, the government said it collected 434 million phone records, down 18% on the year earlier.

But the government came under fire in June 2018 after it emerged the NSA had unlawfully collected 600 million call and text logs without the proper authority. The agency said “technical irregularities” meant it received call detail records it “was not authorized to receive.”

The agency deleted the entire batch of improperly collected records from its systems.

Following the incidents, the NSA reportedly shut down the phone records collection program citing overly burdensome legal requirements imposed on the agency. In January, the agency’s spokesperson said the NSA was “carefully evaluating all aspects” of the program and its future, amid rumors that the agency would not ask Congress to reauthorized its expiring Section 215 powers, set to expire later this year.

In an email Wednesday, the NSA spokesperson didn’t comment on the future of the program, saying only that it was “a deliberative interagency process that will be decided by the Administration.”

The government’s Section 215 powers are expected to be debated by Congress in the coming months.

Detroit-based StockX, which provides a way for people to resell luxury and lifestyle goods including streetwear, bags, watches and shoes, is now valued at over $1 billion based on its most recent raise of $110 million, just revealed by the New York Times. Alongside the raise, StockX is bringing on a new CEO – ecommerce vet and former eBay SVP Scott Cutler.

Cutler replaces co-founder Josh Luber at the helm of the company, but he’ll continue to be the “public face” of the company according to the NYT, which is not unusual for a founder-led company when it brings on more traditionally experienced executives to steer the startup through periods of aggressive growth and business maturation.

StockX’s success rode the sneaker culture boom of the past half-decade or so, as the startup first focused exclusively on acting as a resale source for shoes with high levels of hype. Their unique value prop, for consumers, was offering a verification service so that you knew when you were buying (often at a premium, and often so-called ‘deadstock’ or stuff that’s new in condition but not available through typical consumer sales channels) was the real deal.

The company expanded from there into new categories, first with watches, then handbags, and most recently streetwear – all categories where high potential for fraud mean that consumers are willing to pay more for some assurance of authenticity.

Also unique to StockX is its treatment of the marketplace as analogous to a public stock exchange, with shoe releases, watch, bag and clothing SKUs replacing companies as the trade commodity. The app for StockX displays charts trending value and features bids and calls, making it similar in concept to another company where new CEO Cutler has experience – the NYSE.

With this funding, the company will focus on growing its international business and also do more with selling new products, which it has done on occasion for select releases, but which hasn’t been a primary focus of its business to date.

Many software vulnerabilities are already known, and vendors have even issued patches, but the problem is there are so many patches that it’s often difficult for companies to keep up. Vulcan Cyber wants to help by bringing a level of automation to the patching operation, and in the process reduce exposure to known risks.

Today, it announced a $10 million Series A round from Ten Eleven Ventures and YL Ventures .

In a typical scenario, security researchers find vulnerabilities, the vendors disclose them and patch them. From there it’s up to individual companies to take care of downloading and installing the patch, but Vulcan Cyber co-founder and CEO Yaniv Bar-Dayan says the number of patches has been growing at a furious pace with 6000 patches in 2016, 16,000 in 2017 and 18,000 last year. And that growth trajectory is continuing this year, he says.

Vulcan’s ultimate mission is to help companies remediate security vulnerabilities from their infrastructure. They do this by bringing a level of automation to the process, recognizing that humans can’t keep up with these numbers. “We automate the process of prioritization and deployment to remediate more vulnerabilities faster,” Bar-Dayan explained. What’s more, he said that Vulcan does this without risking business operations, while reducing risk and costs.

Vulcan Cyber risk prioritization view. Screenshot: Vulcan Cyber

The company raised a $4 million seed round last year, bringing the total raised to $14 million so far. As TechCrunch’s Frederic Lardinois pointed out while writing about that seed round, it’s able to achieve this level of automation, while working with the tools developers and security teams typically work with anyway.

“Vulcan Cyber plays nicely with all of the major cloud platforms, as well as tools like Puppet, Chef and Ansible, as well as GitHub and Bitbucket. It also integrates with a number of major security testing tools and vulnerability scanners, including Black Duck, Nessus, Fortify, Tripwire, Checkmarx, Rapid7 and Veracode,” Lardinois wrote.

The company was founded last year and has 25 employees. It plans to continue building its engineering team in Israel with the money from this round, as well as opening an office in San Francisco for sales, marketing and customer success.

Subscription e-book and audiobook service Scribd is unveiling a new content type that it calls a Snapshot — a distillation of a nonfiction book’s key points, which can be read or listened to in 15 minutes or less.

In fact, although it’s only announcing Snapshots today, Scribd says it’s already created around 500 of them, for example, this preview of Jaron Lanier’s audiobook “Ten Arguments for Deleting Your Social Media Accounts Right Now” and this Snapshot of Mark Hanson’s “The Subtle Art of Not Giving A Fuck.”

CEO Trip Adler said a Snapshot is meant to capture “the key insights, the key themes, the tone, the structure of the book.”

As Scribd creates more Snapshots, subscribers will see them show up alongside the books themselves, in both text and audio formats.

And while it’s easy to think of this as a CliffsNotes version that allows you to skip the book itself, Adler said it’s actually designed to do the opposite, namely to help readers “make a better decision” about whether to read the whole thing. In other words, you should get a clear sense of what an author is trying to say, but you’ll have to turn to the book itself if you want the full argument.

In fact, Adler said the idea for Snapshots emerged from a Scribd surveyed of its users about what would make them more likely to read a book.

He also emphasized that Snapshots are “completely original content” created by Scribd’s freelance team — original in the sense that even though they’re citing ideas from the books, they’re not quoting them directly. This puts it in line with Scribd’s original content initiative launched earlier this year, and it’s also an important distinction since the company isn’t striking new deals with publishers or authors to create these previews.

Few debates in technology have been as intense over the past few years as the debate over privacy. As companies like Facebook and Google have swelled with all the minutia of our daily lives — our likes, our searches, our calendar invites — there is increasing awareness of the risks to our privacy as these large data guzzlers suck up our digital selves.

Worse, these companies have used their incumbent network effects to increasingly be the identity layer of the web. From Login with Facebook to being the repository of our data that we shuttle from service to service, it’s increasingly clear that a handful of companies have centralized our data into their own profitable silos.

3Box wants to undo that trend toward identity centralization. Building upon and extending the concepts of the Web3 movement, 3Box offers a decentralized identity cloud storage layer that allows developers to save identity information outside of the big tech stacks.

The company announced today that it has raised a $2.5 million seed round led by Brad Burnham through Placeholder Ventures, where he is a venture partner. Brunham is also a partner at Union Square Ventures, where he invested in developer-focused companies like StackOverflow and Cloudflare. Burnham will join 3Box’s board of directors.

3Box was founded by a trio of alums from ConsenSys, the Ethereum-focused decentralized blob of an accelerator/VC/incubator based in Brooklyn that rode the blockchain wave in 2017 and 2018 only to scale back its ambitions with the decline in asset prices. CEO Michael Sena and COO Danny Zuckerman worked together at ConsenSys company uPort, which was also focused on identity, and they are joined by ConsenSys engineer Joel Thorstensson as CTO.

“We were always interested in the smart identity space … aggregating all of your data and all of your stuff and having agency over that data,” Sena explained. After co-founding uPort though, he increasingly saw an opportunity in distributed databases rather than the mobile application-focused product that uPort was building.

A distributed database is critical for decentralized apps (dApps) to be able to store identity information. “The angle in the Web3 ecosystem that we are seeing, these developers want to write apps with limited backends,” Sena said. “And now they have a distributed place for user data.”

What exactly is identity data though? It “can be anything,” Zuckerman said. “We take the approach of being unopinionated about what kind of data is stored in there.” He added that it “can include data about yourself, browsing data, transaction history, photos.” Ultimately, “identity is the abstraction or emergence of all of that data.”

While there is obviously a privacy and political angle to 3Box’s technology as with many blockchain projects, Sena and Zuckerman see a much more pragmatic reason to use their platform. “It reduces the burden of on-boarding” a new user, Zuckerman said. “Not only can you pull in your profile information, but your other data can also be pulled.” Sena added that “3Box allows you to write a new minimally-viable dApp … our value proposition works best for dApps because they want to avoid the backend.”

While identity layers like 3Box are certainly offered by the big tech giants, other startups have made a play for the space as well. “We’re somewhere between Gravatar and Firebase,” Sena said. “We’re less about owning that [login] button, but enabling the quick on-boarding” of new users.

Developers looking to use 3Box download the company’s JavaScript API library and integrate it into their app. Today, the company offers three API primitives: Profiles (for information like handles and profile photos), Messaging (for communicating between users), and Storage (for app-specific storage).

Right now, 3Box is built around the Ethereum ecosystem. Zuckerman foresees expanding that to other ecosystems as they scale, and the company has also published a roadmap for the rest of 2019 of the new features and themes it intends to focus on.

3Box is co-headquartered in New York City, where Sena and Zuckerman call home, and Berlin, where Thorstensson has his base. With its funding, 3Box intends to build out its team in both locations.

In addition to Placeholder, the round was joined by David Pakman at Venrock, Wendy Xiao Schadeck at Northzone, and Jake Brukhman at CoinFund. The company was funded through ConsenSys’ venture studio, and ConsenSys Global Lead of Media Mike Kriak also sits on the company’s board.

The vision of a universal global ride-hailing service is over. Uber’s decision to exit markets like China, Southeast Asia and Russia coupled with the failure of its rivals to develop a proposed roaming system, means that global travelers must install multiple apps if they are to take advantage of on-demand taxis. That’s unless a little-known startup can turn a bold plan into reality.

In the world of ride-hailing and its billion-dollar investment checks, an $8 million capital raise may not be a big deal but it does represent a coming-out for Splyt, a UK-based startup that is aiming to help make global ride-hailing roaming a reality — and not just within ride-hailing apps.

The four-year-old company announced this week that it closed an $8 million Series A round from a range of undisclosed (and existing) family offices and angel investors. In addition, the round included participation from Southeast Asian ride-hailing company Grab, the firm valued at $14 billion which acquired Uber’s regional business last year.

The deal will see Grab become a Splyt partner and it comes hot-on-the-heels of a similar rollout with Alipay, the digital wallet app run by Alibaba affiliate Ant Financial.

In both cases, Splyt is hooking Alipay and Grab up to its ride-hailing networks to allow users to book (and take) a taxi from another provider within the Alipay or Grab app.

Splyt allows users of Alipay to book taxis on the Grab network in Southeast Asia without downloading Grab’s app

The integration is already live within Alipay for Southeast Asia — Grab is scheduled to work overseas from early 2020 — and it means that users can book and manage rides directly from the payment app thanks to Splyt’s system. In other words, Alipay users can take rides through Grab without having to download the Grab app.

Splyt is not visible to the consumer’s eye. Instead, it lurks behind the scenes acting as the interconnecting services. In that respect, it is much like digital banking services that provide the infrastructure that enables banks to offer digital services. In Splyt’s case, it provides connections for ride-hailing services outside of their markets, but beyond them it allows other apps to access ride-hailing booking features, too.

Relationships are the key part of this offering, beyond Grab and Alipay, Splyt has partnerships with Chinese travel app Ctrip, Careem — the Middle East-based service being acquired by Uber — Gett and car rental service Cartrawler, which added ride-hailing via the tie-up.

“There’s a long way to go to get comfortable with where we are and how close we are to our vision,” Splyt CEO Philipp Mintchin said, admitting that the goal is for all major ride-hailing firms to join.

That said, the existing partner base already gives Splyt reach into some 2,000 cities. The deal with Grab, in particular, will help allow Alipay and Ctrip — two popular services — to open up ride-hailing in Southeast Asia, a region that is an increasingly popular travel destination for Chinese tourists.

Indeed, such is the focus on Asia at this point that Splyt has opened an office in Singapore. Mintchin told TechCrunch that he expects headcount in Singapore will reach 15 this year, mostly on the tech side, while overall the company is predicted to grow to 50 people by the end of this year.

“Most of our business and partners are based out of Asia,” he added of the new office.

The Splyt Team at the company’s office in London

While connecting ride-hailing services and popular apps makes absolute sense for consumers who can enjoy the convenience of roaming, navigating and securing partnerships is not straightforward in today’s ride-hailing world. Aside from a network of complicated relationships — Uber and Didi, in particular, are investors in many competing services and each other — many companies are also developing new features behind simply taxis.

Mintchin declined to discuss potential deals but he did tease that Splyt is working to onboard a number of new partners this year.

“In this industry, everyone is talking to everyone,” he said of the partnership push.

Mintchin admitted that the “politics of the ride-hailing industry” mean that some companies refuse to work with others — no names named, alas — and others prefer to work with specific firms, too. Then there’s also an element of trust involved with giving a third party access to a service which ends up being used by yet another third party.

“We are here to partner and benefit each other rather than to try to steal a fleet and run our own app,” he said of Splyt’s neutral position and its role as the behind-the-scenes integrator. “We are not all of a sudden going to influence the partners we work with… the partners make decisions.”

It’s a patient game, but already Splyt is seeing growth double on a weekly basis since May. In some areas, Mintchin said that the service is seeing a 90 percent repeat use through its partners. Going forward, he added, the Series A funding will go towards closing those supply gaps to make the service more usable in more locations.

It’s an audacious vision but, given the balkanization of the industry in recent years, it remains the best hope that travelers have of delivering on the vision of using their favorite ride-hailing app anywhere in the world.