Skip to main content

Facebook is learning that hardware is hard

Facebook is learning that hardware is hard

/

Early sales of the Portal has been slow — but the company can’t afford to quit now

Share this story

Photo by Nick Statt / The Verge

Meet me at our first-ever Interface Live on Tuesday, October 22nd! I’ll be chatting with misinformation researcher Renee DiResta. Get tickets.

One of the first adages I heard as a tech reporter is a three-word refrain uttered by anyone who’s ever sought to make something with atoms rather than bits: Hardware is hard. Make a mistake in software and you can push out a fix more or less immediately; make a mistake with hardware and the phones explode. In the old days, a tech company typically made software or hardware, but these days that or has been replaced with an and. After Apple’s outsized success with the iPhone, all the other tech giants got religion on the value of software and hardware working together. And suddenly companies like Microsoft, Google, and Facebook, who once limited themselves to high-margin software businesses, had to invest untold billions in figuring out to make and sell physical objects.

Facebook first dipped a toe in these waters when it bought Oculus, maker of virtual reality hardware, in 2014. Since then, as Salvador Rodriguez recounts in CNBC today, the company has struggled to gain traction in hardware markets. It spun up a research and development lab called Building 8 in 2016, making sci-fi promises about technologies that would let us hear with our skin and type with our minds. But Building 8 was reorganized into nonexistence just two years later — with its only lasting contribution to Facebook’s product lineup being the video-chat device Portal.

Portal launched last year to a tech press skeptical that consumers would welcome into their homes a camera and microphone from a company that has spent the past few years battered by data privacy scandals. The skepticism appears to have been warranted, Rodriguez reports:

It’s unclear how the new Portal products will fare, but the first generation released in 2018 shipped only 54,000 units, according to IDC. Facebook disputes that estimate, but the company has never offered its own figure and declined to do so at the second-generation Portal unveiling on Tuesday.

(Another estimate puts the figure at somewhere under 300,000 units, or less than 1 percent of the market.)

At the same time, it can take a brand years to build market share. There was never a question in my mind whether Facebook would build a second generation of Portal devices — only what they would look like when they arrived, and how the world would react when they did.

On Wednesday we found out. The company introduced three new version of Portal: an 8-inch and a 10-inch version that look like less functional iPads, and an accessory called Portal TV that plugs into your television. As before, a major theme of coverage was the privacy risks associated with the device. And analysts remain fairly bearish on Portal in the near term, Heather Kelly reports in the Washington Post:

Portals will account for just 4 percent of smart-display shipments in 2019, says David Watkins, an analyst at market research firm Strategy Analytics. He cites high prices and privacy concerns as reasons the smart displays haven’t taken off more.

The most popular smart display in North America is Google’s Nest Hub, with Amazon’s Echo Show in second place, according Strategy Analytics. It estimates the smart-display market will hit 31 million devices globally in 2019. (The devices are especially popular in China, where companies like Baidu and Xiaomi sell inexpensive versions of their own.)

So why keep going?

Andrew “Boz” Bosworth, the longtime Facebook executive who now runs the hardware division, says Portal is a quintessential Facebook product. “This product is the very core essence of what Facebook does,” he said Tuesday at a product demonstration I attended in San Francisco. “It connects you meaningfully with the people you care most about.”

The way it connects people is largely through video calls, which it spiffs up with augmented reality features, a camera that tracks people around the room, and other enhancements. To me it all feels like a lot of work to make video calls marginally better, but Bosworth told me that Portal is something different.

“Video calling on the phone always felt very transactional,” he told me. “Your left hand is sacrificed to the call until you’re done. [You’d think] ‘how can I get off this call as politely as possible? Whereas Portal is a lot more like hanging out. You don’t have to be doing anything. That, I think, is something that is interesting.”

Sure, I said. But why not just buy an iPad and use Skype or FaceTime?

“Right now those are not good experiences,” Bosworth said. “Go set an iPad up and spend a ton of time video-calling people you care about. See how enjoyable it, is or how annoying it is. And then do it with Portal.

“This is where I think people go wrong,” he continued. “People just look at the features on the box … That’s not where this lives. This lives in the experience.”

It remains to be seen how many more people will sign up for that experience than did last year. In the meantime, it’s clear that the story over privacy and in-home displays doesn’t begin and end with Facebook. On the same day news of Portal came out, the Washington Post reported on how your smart TV is monitoring what you watch to target ads at you, and the Financial Times reported that your smart TV is sending market research data to Netflix and personally identifying information to Google.

The Ratio

Today in news that could shape public perception.

🔼 Trending up: Google and Facebook might begin paying publishers directly to license their content, as antitrust probes into their allegedly monopolistic business practices move forward. I’ll take it!

🔼 Trending up: Twitter is dismantling an effort to discredit Hong Kong protesters that likely originated in China, and Facebook and YouTube followed it.

🔽 Trending down: Facebook faces new charges from a human rights nonprofit that its ad engine was used for housing discrimination.

Oversight

A couple follow-up notes from yesterday’s column about Facebook’s forthcoming judicial branch. One, I’m told that the current working name is the “oversight board,” rather than the “independent oversight board,” even though that language was in yesterday’s blog post. The board will eventually be able to amend its bylines to change the name if it wants, Facebook told me.

Of more consequence: Evelyn Douek has a good post at Lawfare in which she explores how a recent update to Facebook’s values could affect how the board ultimately makes its decisions. Notably — if inevitably — the new values weigh the importance of “voice” above that of “safety,” she writes:

Therefore, in cases of ambiguity, when the extent of risk posed by a certain category of speech is not necessarily clear, Facebook will not—and, if the Oversight Board project works, cannot—err on the side of caution and simply take that content down “just in case.” This reflects a certain degree of risk tolerance on Facebook’s part—which will no doubt be praised by those committed to a robust marketplace of ideas. It also perhaps reflects the changing role of these private companies that facilitate so much public discourse: They are not literally, technically or legally a new public square, but their systemic importance means that society is coming to have expectations of their responsibility to the public that go beyond mere consumer satisfaction.

Douek also had a nice gloss on yesterday’s big charter reveal at The Atlantic. Yes to all this:

The Oversight Board is fundamentally a bet by Facebook that the legitimacy of its decisions matters—and matters more than getting its way on every question. As other platforms seem to double down on the idea that they do not need to publicly explain their decisions or abide by their own rules when it does not suit their short-term interests, Facebook appears to be making a different wager: Accountability and legitimacy can reassure users—and regulators—of the value of its product. Acting in bad faith would undermine Facebook’s own gamble. Legitimacy, Facebook hopes, will become part of its value proposition.

Governing

Congress is drafting a bill to create a national commission that will study the ways social media can be weaponized. The commission will also assess how effective tech companies are in protecting users from harmful content. The news comes after a hearing in which members questioned Google, Facebook, and Twitter about the connection between social networks and real-world violence. Tony Romm and Drew Harwell at The Washington Post:

The draft House bill obtained by The Washington Post is slated to be introduced and considered next week. If passed, the commission would be empowered — with the authority to hold hearings and issue subpoenas — to study the way social media companies police the Web and to recommend potential legislation. It also would create a federal social media task force to coordinate the government’s response to security issues.

People with Amazon Alexa speakers can start making campaign donations to 2020 presidential candidates using their devices starting tomorrow. But the rules say you can only donate to “principal” candidates, and Amazon hasn’t yet clearly defined who falls into this category. (Makena Kelly / The Verge)

At least 75 countries, including the U.S. and Germany, have adopted China’s surveillance strategies, using facial recognition software to track citizens. Much of the software comes from Chinese telecom company Huawei. (Ryan Tracy / The Wall Street Journal)

The Chinese government ran a Russian-style disinformation campaign on Twitter in an effort to discredit Hong Kong protestors. Twitter has taken down 1,000 fake accounts linked to this operation, and suspended 200,000 more. (Raymond Zhong, Steven Lee Myers, and Jin Wu / The New York Times)

Hong Kong protesters make clever use of tools including Telegram, Twitter, and live maps. The entrepreneur and social critic Maciej Cegłowski shares characteristically sharp notes . (Maciej Cegłowski / Idle Words)

London’s Met police force is teaming up with Facebook in an effort to prevent live streaming terrorist attacks around the world. The Met will give Facebook video footage from officer trainings to help the company develop technology that can detect when someone is live-streaming an attack. (Metropolitan Police)

Industry

Amazon is tightening its control over the third-party apps used by many businesses to sell products on the platform. The apps have access to customer data, and some have been violating Amazon’s privacy policy by advertising to customers on Facebook. Louise Matsakis report at Wired:

To streamline their operations, many sellers rely on specialized business apps that tap into the Amazon Marketplace Web Service APIs, which can integrate data including sensitive customer information like names, emails, and delivery addresses. There are tools that automate simple tasks, like printing shipping labels, as well as apps that monitor key metrics like user reviews and sales volumes, which determine whether products appear higher in Amazon’s search results—the most popular way to shop on the site. While Amazon has multiple policies governing the use of these apps and their data, the cottage industry that sprung up around Amazon MWS has been relatively decentralized. Amazon only launched its Marketplace Appstore in May 2018.

Now Amazon is cracking down on third-party apps that accessed customer information through MWS and are in violation of its policies. Earlier this year, the company began emailing developers that they had to submit information about their apps in order to continue using Amazon MWS. Seller forums are full of posters wondering when they’ll finally regain access. WIRED spoke with three Amazon developers who received warnings or had their API access revoked in recent months. One enabled Amazon sellers to create targeted advertisements on Facebook using customer data for over a year, in violation of Amazon’s privacy policy.

Instagram is cracking down on posts related to diet products and cosmetic surgery. Under the new rules, posts promoting weight-loss products or miracle cures will be hidden from users under the age of 18 — or removed altogether. (PA Media / The Guardian)

Instagram influencers making money by charging fans to see their “Close Friends” posts. It’s unclear what Instagram itself thinks of the practice, though among other things it illustrates how few legitimate ways that the platform’s stars currently have to make money there. (Kaitlyn Tiffany / The Atlantic)

Facebook is pouring money into AI research to teach chatbots to converse like humans. Facebook’s virtual assistant M may have flopped, but it continues to invest heavily in chatbot research. (Mark Sullivan / Fast Company)

After a CNBC article yesterday described Facebook’s plans to develop augmented-reality glasses, Alex Heath reports there are actually two projects underway. The company’s reported partnership with Ray-Ban is for glasses similar to Snap’s Spectacles, he writes. (Alex Heath / The Information)

A report from Data and Society raises doubts that AI can fix deepfakes. Facebook has recently been promoting AI solutions to doctored videos, but the report argues it won’t be enough. (Zoe Schiffer / The Verge)

Youtube added a big giant advertisement at the top of its TV app. The desktop site site already has ads in this spot, and it’s considered premium real estate for advertisers. (Julia Alexander / The Verge)

An inside look at the history behind Google’s race to become the leader in AI, featuring interviews with CEO Sundar Pichai and AI chief Jeff Dean. The company’s next big project: quantum computing. (Katrina Brooker / Fast Company)

Google’s screen time manager for parents, Family Link, added new features that allow them to limit screen time for individual apps, instead of the device as a whole. It also allows people to extend screen time as needed. (Sarah Perez / TechCrunch)

LinkedIn introduced Skills Assessments: short, multiple-choice tests to let users show off their computer science know-how, along with other work-related skills, to potential employers. (Ingrid Lunden / TechCrunch)

Fortnite added cross-platform voice chat — another step toward making the popular free-to-play game a full-fledged social network. It’s based on a similar feature developed by Houseparty, the group video chat app that Fortnite maker Epic acquired in June. (Casey Newton / The Verge)

And finally...

They Are What You Eat

Caity Weaver visits the headquarters of your favorite mall food court brands and, among other things, chats with their social media teams:

“If you jump in inappropriately, you will get roasted,” he said, reflecting on the recent Twitter storm over Popeyes’ Spicy Chicken Sandwich. “Like Zaxby’s jumped in, they got roasted. Boston Market jumped in. Chick-fil-A took a lot of L’s.”

Mr. Ayala’s job is, essentially, to talk about Moe’s in a brief, hilarious and charming way, without stopping, forever. He found delirious success one day this summer when his tweet combining a meme about aliens in Area 51 with the very notion of Moe’s burritos received roughly 2,100 retweets. But then he had to tweet again.

Relatable as hell, honestly.

Talk to us

Send us tips, comments, questions, or call us from your Facebook Portal: casey@theverge.com and zoe@theverge.com.

Subscribe to the newsletter

Get The Interface delivered to your inbox each evening by subscribing here.