Year: 2018

Though Facebook’s focus on Day 1 of its F8 conference centered on the company’s recent struggles and their relationship with the phrase “taking broader responsibility,” Day 2 shifted most of the pizazz to the technical advances its giant team has made over the past year.

Today, the company announced PyTorch 1.0, a new iteration of the framework that merges Python-based PyTorch with Caffe2 allowing developers to move from research to production in a more frictionless way without having to deal with migration.

At Facebook, the company’s AI efforts are split between two teams, the Facebook AI Research group (FAIR) and the company’s Applied Machine Learning team (AML). The distinction ultimately boils down to one division researching AI with seemingly limitless computational resources at their disposal and the other looking to implement lightweight machine learning models more suited for consumers. In the past, the former mission has been better-suited for the research-optimized PyTorch while the more resource-efficient Caffe2 framework has made sense for the latter.

Moving between these frameworks through research and production phases has been a bit of a headache for Facebook. PyTorch 1.0 fuses the best of both worlds with immediate and graph execution modes that should continue its research-focused prowess while gaining some optimizations for production.

“Instead of needing a framework for training/research and a framework for production, now you only need to use PyTorch 1.0 and you can seamlessly go from research to production in one framework,” a Facebook spokesperson told TechCrunch.

Facebook collaborated with Microsoft last year to create the Open Neural Network Exchange (ONNX) which was designed to make movements between frameworks more simple. They’ve now integrated ONNX into PyTorch 1.0 so that the models can be interoperable with other frameworks and developers can “mix-and-match.”

The company also says that they will be unifying the PyTorch 0.4 and Caffe2 codebases over the next several months, to create a unified framework that supports several features including efficient graph-mode execution with profiling, mobile deployment and extensive vendor integrations.

PyTorch 1.0 will be released in beta later this year. The company says that Microsoft is planning to support the framework in Azure and that Amazon will support PyTorch 1.0 across their cloud products.

The Handmaid’s Tale, America’s favorite dystopian drama, will return for a third season.

Hulu announced the news this morning on the heels of revealing it has hit 20 million subscribers for its streaming service.

The second season of The Handmaid’s Tale is currently airing on Hulu, with the third episode to available today.

The series is based on Margaret Atwood’s 1985 novel by the same name. In it, humanity has become mostly infertile, leading a conservative militia to overtake the country and enslave the remaining fertile women as childbearers for the “upstanding” families of the regime.

In Season One, the story was focused on exposition of the new world as well as our main character, Offred (June) and her life as a Handmaid. Season 2 seems to be expanding on that world and our main character, showing both the good and the bad of her character’s personality.

The Handmaid’s Tale is far and away Hulu’s most popular and critically acclaimed original series, winning five Emmys last year, including one for Best Drama series.

If you want to learn more about The Handmaid’s Tale, make sure to check out the latest episode of TC’s Original Content podcast.

It seems like Hulu and Spotify had the right idea, when they launched a discounted package last month that combined both of their services. Now, the low-cost streaming service Philo is doing the same – but not with Spotify. Instead, Philo and Pandora are teaming up on a combo deal of their own: $16 per month for 37 Philo channels and Pandora Premium. Or, for $20 per month, you’ll get 42 channels and Pandora Premium.

Oh, but there’s a catch: Pandora Premium is only free for three months, not for the length of your subscription. Bummer.

That’s a different deal than what Hulu and Spotify are offering, and is definitely more of a promo than a partnership.

Still, it’s interesting to see others in the industry using the video/music combo bundle to sell subscriptions.

Philo is a relative newcomer to the streaming TV market.

Though the company had been around for some time, offering an on-campus internet TV service for universities, its $16 per month streaming TV service only launched this past November. (Philo can get away with the lower pricing because it ditches what makes streaming so expensive: sports.)

Instead, its core lineup includes cable networks like A&E, AMC, Animal Planet, AXS TV, BBC America, BET, Cheddar, CMT, Comedy Central, Discovery Channel, DIY, Food Network, FYI, GSN, HGTV, HISTORY, Investigation Discovery (ID), IFC, Lifetime, MTV, MTV2, Nickelodeon (+Nick Jr., Teennick), OWN, Paramount Network, Science Channel, TLC, Travel Channel, VH1, VICELAND, We, and others.

The 46-channel package adds American Heroes Network, BET Her, Cooking Channel, Destination America, Discovery Family, Discovery Life, Logo, MTV Live, and NickToons.

Philo’s service include a 30-day DVR, the ability to watch shows that aired in the past three days, and the ability to pause live TV.

Meanwhile, Pandora Premium (basically, Pandora’s Spotify competitor) is typically $9.99 per month. It usually offers a free trial, but currently that’s 60 days, not three months.

In other words, the Philo combo deal means you get a longer free trial of Pandora Premium along with your Philo subscription. Of course, Pandora’s goal is to then convert those free trial users to paid subscribers.

The move to offer a bundled three-month free trial of Pandora Premium with Philo comes shortly after rival Apple Music began extending its own three-month free trial by another month to those who hadn’t yet signed up, according to reports. And it comes just ahead of Spotify’s earnings, which investors seem optimistic about.

Facebook is hosting its F8 developer conference in San Jose this week. Yesterday, it kicked off with a keynote from CEO Mark Zuckerberg, who addressed the company’s latest scandals, including Cambridge Analytica, election meddling, and fake news. But he also spoke of his commitment to Facebook’s mission of connecting the world, and his optimism for the future. The company then introduced a number of new products across its platform, as well as on Instagram, WhatsApp, and Oculus.

We saw things like Facebook Dating, Instagram video calling, Oculus Go’s launch, a new Messenger, VR memories, 3D Photos, and more.

Today, live coverage of F8 2018’s Day 2 begins again at 10 AM PT, 1 PM ET, Wednesday May 2.

On the schedule is a second, less lengthy keynote address.

Typically the Day 2 keynote focuses more on research, like machine learning technology, and other forward-looking projects and developments. We may hear from Facebook’s research divisions and perhaps the Telecom Infra Project. (Details on the keynote’s speakers haven’t yet been provided, so these are just guesses for now.)

The full F8 schedule is here.

The keynote will stream to Facebook’s F8 website, though yesterday it experienced a lot of glitches. Even the F8 website was down for a while. If that happens again, you can just follow along the news via TC’s Twitter account or here on the site.

TechCrunch is coming to Tel Aviv, Israel to unpack the revolution that is coming in Mobility. So we’re excited to announce that a pioneer of the space, Austin Russell, will be speaking.

Austin Russell is an accomplished applied physicist, and founder and CEO of Luminar Technologies, Inc. He founded the company in 2012 with a vision to develop a new type of LiDAR for the autonomous vehicle industry.

Austin began his career as an inventor at an early age, writing his first patent by the time he was 12. A year later, his interests became focused on the photonics industry, where he later developed a number of projects – ranging from laser-based wireless power transmission to projected augmented reality systems. Many of his projects required the creation of custom 3D mapping sensors. He made an early bet on the application of high-performance LiDAR for the autonomous vehicle industry with the foresight that it would transform the future of transportation.

In 2012, Austin founded Luminar while performing independent research at the Beckman Laser Institute. He was subsequently recruited to Stanford to study in the applied physics department but dropped out 3 months later to accept the Thiel Fellowship. Since then, the Luminar team has achieved a number of breakthroughs in LiDAR, which can finally make autonomous vehicles both safe and ubiquitous.

You need to come to this event!

Tickets are available here.

It’s hard to keep track with what MoviePass is doing these days, but here’s trying. Today, MoviePass has brought back its nearly all-you-can-watch monthly movie subscription, Variety reports. With its original plan, subscribers can watch up to one movie a day, every day.

This comes after MoviePass removed this subscription last month in favor for a promotional $29.95 three-month plan that only let people see four movies a month. At the time, however, MoviePass CEO Mitch Lowe said the move didn’t mean the unlimited plan would never come back.

“We’re continually testing various promotions with different partners, and the current iHeartRadio deal is consistent with that approach,” a MoviePass spokesperson said at the time. “This does not mean that our unlimited subscription will not be offered in the future.”

MoviePass currently offers two plans — the $9.95 a month “unlimited” one, and a $7.95 a month plan that lets you see just three movies per month and includes a three-month trial to iHeartRadio’s all-access plan.

But there’s another service on the market that’s caught my eye as of late. It’s called Sinemia, which costs $15.99 a month for three movies per month. Yes, it’s more expensive than MoviePass but the key differentiator is the fact that you can buy movie tickets online, ahead of time. With MoviePass, you have to physically go to the theater to purchase your tickets.

I’ve reached out to MoviePass and will update this story if I hear back.

The massage-on-demand service Soothe seems to be rubbing investors the right way with the close of a new $31 million round of funding.

The Series C round from late stage and growth capital investment firm, The Riverside Company, caps a busy first quarter for the massage service. It also relocated from Los Angeles to Las Vegas; named a new chief executive; and announced new geographies where its massage booking platform is now available. 

As part of the new round, chief executive and founder Merlin Kauffman is stepping down from the role and assuming the mantle of executive chairman. Current chief financial officer, Simon Heyrick, is stepping into the chief executive role.

The former CFO of MarketShare, Heyrick has helped the company expand to over 11,000 massage therapists in its network.

The company said the new round would help keep massage therapists in its network with pricing that can be up to three times more than those therapists would make in their local markets.

Beyond the new financing and a new boss, Soothe is also heading to new markets, launching services in Manchester, UK; Australia’s Gold Coast, PIttsburgh, and Hartford, Conn. (some of those places are not like the others).

Soothe isn’t the only player in the massage marketplace. New York-based Zeel also has an offering for folks who want to book massages on the fly. Zeel claims a geographic reach of 85 U.S. cities while Soothe claims roughly 60 cities worldwide.

SALT Lending has issued $40 million in asset-backed loans that are aimed at investors in cryptocurrencies to give them a bit of liquidity without the need to sell off assets. 

The company rolled out its first loans late in 2017 after two years spent developing the technology to make the loans work. SALT Lending relies on a multi-signature wallet that allows counter-parties in a transaction to both access the account — and a software service that marks to market the value of the cryptocurrencies held as assets.

Founded by Blake Cohen, a former real estate executive for his family’s own business, SALT Lending is Cohen’s first foray into cryptocurrency entrepreneurship.

Cohen says the idea for the service was a no-brainer. “Cryptocurrencies and blockchain assets make the ideal collateral,” says Cohen. In typical asset-backed lending the assets need to be located, seized and liquidated for lenders to be made whole. With cryptocurrencies all of those problems go away, says Cohen.

Most of the capital that SALT is lending has come from the company’s own assets — and through SALT Blockchain Asset Management, which is a fund with several individual limited partners.

Currently the company is servicing individual borrowers, but Cohen foresees a time when exchanges, miners, and large tokenized startups can gain additional liquidity through the collateralized loans.

“The current market cap [for cryptocurrencies] is $400 billion and that is totally unleveraged. That is a gigantic inefficiency in the market,” Cohen says.

The company currently offers 36 month loans and (depending on the loan to value ratio) the company will loan up to 60% of the value of the cryptocurrency collateral, says Cohen. Interest rates on the loans range from 12% to 22%, Cohn says. SALT Lending is only lending against Ethereum and Bitcoin for the moment.

“Basically the money that we put in is as a proof of concept to entice the institutional capital providers to get involved with this,” Cohen says. “This point is to one, demonstrate that there’s a market for this, and two, that you can do this and make risk palatable to the lender.”

Google is betting big on the Assistant ecosystem and it’s now putting its money where its mouth is. As the company announced today, it’s launching a new program that will provide investment capital and other resources to early-stage startups that build applications in the Google Assistant ecosystem.

Typically, companies announce these kind of program to kickstart an ecosystem around a new product. While developers have already launched plenty of services for the Google Assistant, the company says that it is launching this new program to “promote more of this creativity.”

Google VP for search and the Google Assistant Nick Fox echoed this. “With the Google Assistant, we’re focused on fostering an open ecosystem for developers, device makers, and content partners to build new experiences,” he told me. “We’re already seeing a lot of creativity from developers with the Google Assistant, and to help promote this, we’re opening a new investment program for early-stage startups.”

Investments are one part of this program, but Google will also work with these startups directly to provide them mentorship and advice from engineers, product managers and design experts. The startups in the program will also get early access to new features and tools, as well as access to the Google Cloud Platform and promotional support. That sounds a bit like an accelerator program, though that’s not quite what Google is calling it.

Fox tells me that Google won’t put a cap on the investments. “We’ll invest as much as we see fit, and are focused on helping startups succeed in this emerging space,” he said. “And we’re not just offering investment capital. We’re eager to partner with these startups and leverage our company’s strengths to help these products come to market poised for success.”

The first startups in this program include GoMoment, a concierge service for hotels and Edwin, a personal English tutor, as well as the developer tools BotSociety and Pulse Labs.

These startups are good representatives of what Google is after. Fox tells me that Google is looking for startups that are “pursuing an interesting space for Assistants, such as vertical industries like travel or gaming.” In addition, Google is also looking to deepen some of its partnerships, but for the most part, it’s simply looking for startups that are pushing technologies like the Assistant forward.

While the mid 2010s were filled with massive financing rounds for companies looking to shift business models like Instacart, Uber, and others along those lines, today’s hottest sector is a more technical one: artificial intelligence.

That’s required a different approach to building companies. As tools become more and more powerful, and frameworks like TensorFlow become more robust, building a company centered around AI — and the big technical problems that either feed it or stem from it — are powering the next wave of potentially massive startups. That’s why Lu Zhang and Homan Yuen, two Stanford technical graduates, are raising a big fund based on their extensive technical experience to find companies that have that strong technology backbone that will become the next big startup.

The pair are raising $85 million for a new fund called Fusion Fund. Fusion Fund’s companies are going to be focused on connected industries, networking technology like communications protocols, data-rich AI products and some health and medical devices. That third bit — investing in AI — is what pretty much every fund is doing, though Yuen made it clear the company isn’t investing in just algorithms, which are eventually going to be a race to the bottom.

“We noticed there weren’t many venture capitalists focusing on technology companies and actually have technical backgrounds,” Yuen said. “We thought, that’s a good opportunity to help early stage companies that need different help, and advice, and connections.”

Zhang and Yuen are two technical founders that sold their companies and were looking to figure out what to do after that. In Zhang’s case, with a masters in material science, she was approached by investors to do a fund, while Yuen had known her at Stanford and ended up syncing up to work on the fund, he said. The fund will be focused on seed-stage companies, looking to write checks between $500,000 and $1 million for companies with technical founders and a pretty difficult problem to solve that requires that expertise.

“We’re not funding science products, or expedition missions, we want clear vision in building a business on the technical advantages,” Yuen said. “Part of our thesis is, over the last five to ten years, we saw a lot of investment into business model innovation. We decided it was time to flip back to tech and infrastructure investing.”

On that front, a lot of these areas are going to get a lot more interesting as time goes on. The tech industry is in the early stages of 5G development and rollout, and once there are more rigorous standards, there will likely be a lot of activity in that space, such as new connection protocols. There’s also a strong security element to that, and all of this requires heavy technical expertise to build something pretty defensible, Yuen said.

But if you want just one example — and this is one where Yuen is sitting on the sidelines for now — on how technical these problems are getting, you can look at the rapid emergence of the custom AI chip space. That problem requires an expertise in rethinking the actual silicon where operations happen to more efficiently tackle machine learning problems, whether that’s focusing on speed, lowering the power consumption, reducing the overall space the processors take up, or getting the cost down to something more reasonable than a bleeding-edge GPU. Right now the space has a ton of VC money flowing into it, and given the size of the checks Yuen is looking to write, it’s an area where the firm has to be more thoughtful about the companies it picks.

This is actually Fusion Fund’s second fund, as it previously went under a different name. First called NewGen Capital, Zhang and Yuen raised $17 million in its first fund and backed 36 seed-stage companies in companies like TVision Insights, Stratifyd, Paperspace, MissionBio, and Paradromics.