In May, Google announced the launch of a new website aimed at explaining how they serve and address news across Google properties and platforms.
The site, How News Works, states Google’s mission as it relates to disseminating news in a non-biased manner. The site aggregates a variety of information about how Google crawls, indexes, and ranks news stories as well as how news can be personalized for the end user.
How News Works provides links to various resources within the Google news ecosystem all in one place and is part of The Google News Initiative.
What is The Google News Initiative?
The Google News Initiative (GNI) is Google’s effort to work with news industry professionals to “help journalism thrive in the digital age.” The GNI is driven and summarized by the GNI website which provides information about a variety of initiatives and approaches within Google including:
- How to work with Google (e.g., partnership opportunities, training tools, funding opportunities)
- A list of current partnerships and case studies
- A collection of programs and funding opportunities for journalists and news organizations
- A catalog of Google products relevant to journalists
Google attempts to work with the news industry in a variety of ways. For example, it provides funding opportunities to help journalists from around the world.
Google is now accepting applications (through mid-July) from North American and Latin American applicants to help fund projects that “drive digital innovation and develop new business models.” Applicants who meet Google’s specified criteria (and are selected) will be awarded up to $ 300,000 in funding (for U.S. applicants) or $ 250,000 (for Latin American applicants) with an additional award of up to 70% of the total project cost.
The GNI website also provides users with a variety of training resources and tools. Journalists can learn how to partner with Google to test and deploy new technologies such as the Washington Post’s participation in Google’s AMP Program (accelerated mobile pages).
AMP is an open source initiative that Google launched in February 2016 with the goal of making mobile web pages faster.
AMP mirrors content on traditional web pages, but uses AMP HTML, an open source format architected in an ultra-light way to reduce latency for readers.
News transparency and accountability
The GNI’s How It Works website reinforces Google’s mission to “elevate trustworthy information.” The site explains how the news algorithm works and links to Google’s news content policies.
The content policy covers Google’s approach to accountability and transparency, its requirements for paid or promotional material, copyright, restricted content, privacy/personalization and more.
This new GNI resource, a subsection of the main GNI website, acts as a starting point for journalists and news organizations to delve into Google’s vast news infrastructure including video news on YouTube.
Since it can be difficult to ascertain if news is trustworthy and accurate, this latest initiative by Google is one way that journalists (and the general public) can gain an understanding of how news is elevated and indexed on Google properties.
The post Google’s How News Works, aimed at clarifying news transparency appeared first on Search Engine Watch.
Voice recognition is a standard part of the smartphone package these days, and a corresponding part is the delay while you wait for Siri, Alexa or Google to return your query, either correctly interpreted or horribly mangled. Google’s latest speech recognition works entirely offline, eliminating that delay altogether — though of course mangling is still an option.
The delay occurs because your voice, or some data derived from it anyway, has to travel from your phone to the servers of whoever operates the service, where it is analyzed and sent back a short time later. This can take anywhere from a handful of milliseconds to multiple entire seconds (what a nightmare!), or longer if your packets get lost in the ether.
Why not just do the voice recognition on the device? There’s nothing these companies would like more, but turning voice into text on the order of milliseconds takes quite a bit of computing power. It’s not just about hearing a sound and writing a word — understanding what someone is saying word by word involves a whole lot of context about language and intention.
Your phone could do it, for sure, but it wouldn’t be much faster than sending it off to the cloud, and it would eat up your battery. But steady advancements in the field have made it plausible to do so, and Google’s latest product makes it available to anyone with a Pixel.
Google’s work on the topic, documented in a paper here, built on previous advances to create a model small and efficient enough to fit on a phone (it’s 80 megabytes, if you’re curious), but capable of hearing and transcribing speech as you say it. No need to wait until you’ve finished a sentence to think whether you meant “their” or “there” — it figures it out on the fly.
So what’s the catch? Well, it only works in Gboard, Google’s keyboard app, and it only works on Pixels, and it only works in American English. So in a way this is just kind of a stress test for the real thing.
“Given the trends in the industry, with the convergence of specialized hardware and algorithmic improvements, we are hopeful that the techniques presented here can soon be adopted in more languages and across broader domains of application,” writes Google, as if it is the trends that need to do the hard work of localization.
Making speech recognition more responsive, and to have it work offline, is a nice development. But it’s sort of funny considering hardly any of Google’s other products work offline. Are you going to dictate into a shared document while you’re offline? Write an email? Ask for a conversion between liters and cups? You’re going to need a connection for that! Of course this will also be better on slow and spotty connections, but you have to admit it’s a little ironic.
Facebook secretly retracted messages sent by CEO Mark Zuckerberg, TechCrunch reported seven months ago. Now for the first time, Facebook Messenger users will get the power to unsend too so they can remove their sent messages from the recipient’s inbox. Messages can only be unsent for the first 10 minutes after they’re delivered so that you can correct a mistake or remove something you accidentally pushed, but you won’t be able to edit ancient history. Formally known as “Remove for Everyone,” the button also leaves a “tombstone” indicating a message was retracted. And to prevent bullies from using the feature to cover their tracks, Facebook will retain unsent messages for a short period of time so if they’re reported, it can review them for policy violations.
The Remove feature rolls out in Poland, Bolivia, Colombia and Lithuania today on Messenger for iOS and Android. A Facebook spokesperson tells me the plan is to roll it out globally as soon as possible, though that may be influenced by the holiday App Store update cut-off. In the meantime, it’s also working on more unsend features, potentially including the ability to preemptively set an expiration date for specific messages or entire threads.
“The pros are that users want to be in control . . . and if you make a mistake you can correct it. There are a lot of legitimate use cases out there that we wanted to enable,” Facebook’s head of Messenger Stan Chudnovsky tells me in an exclusive interview. But conversely, he says, “We need to make sure we don’t open up any new venues for bullying. We need to make sure people aren’t sending you bad messages and then removing them because if you report them and the messages aren’t there we can’t do anything.”
Zuckerberg did it; soon you can, too
Facebook first informed TechCrunch it would build an unsend feature back in April after I reported that six sources told me some of Mark Zuckerberg’s Facebook messages had been silently removed from the inboxes of recipients, including non-employees with no tombstone left in their place. We saw that as a violation of user trust and an abuse of the company’s power, given the public had no way to unsend their own messages.
Facebook claimed this was to protect the privacy of its executives and the company’s trade secrets, telling me that “After Sony Pictures’ emails were hacked in 2014 we made a number of changes to protect our executives’ communications. These included limiting the retention period for Mark’s messages in Messenger.” But it seems likely that Facebook also wanted to avoid another embarrassing situation like when Zuckerberg’s old instant messages from 2004 leaked. One damning exchange saw Zuckerberg tell a friend “if you ever need info about anyone at harvard . . . just ask . . . i have over 4000 emails, pictures, addresses, sns.” “what!? how’d you manage that one?” the friend replied. “People just submitted it . . i don’t know why . . . they ‘trust me’ . . . dumb fucks” Zuckerberg replied.
The company told me it was actually already working on an Unsend button for everyone, and wouldn’t delete any more executives’ messages until it launched. Chudnovsky tells me he felt like “I wish we launched this sooner” when the news broke. But then six months went by without progress or comment from Facebook before TechCrunch broke the news that tipster Jane Manchun Wong had spotted Facebook prototyping the Remove feature. Then a week ago, Facebook Messenger’s App Store release notes accidentally mentioned that a 10-minute Unsend button was coming soon.
So why the seven-month wait? Especially given Instagram already allows users to unsend messages no matter how old? “The reason why it took so long is because on the server side, it’s actually much harder. All the messages are stored on the server, and that goes into the core transportation layer of our how our messaging system was built,” Chudnovsky explains. “It was hard to do given how we were architected, but we were always worried about the integrity concerns it would open up.” Now the company is confident it’s surmounted the engineering challenge to ensure an Unsent message reliably disappears from the recipient.
“The question becomes ‘who owns that message?’ Before that message is delivered to your Messenger app, it belongs to me. But when it actually arrives, it probably belongs to both of us,” Chudnovsky pontificates.
How Facebook Messenger’s “Remove for Everyone” button works
Facebook settled on the ability to let you remove any kind of message — including text, group chats, photos, videos, links and more — within 10 minutes of sending. You can still delete any message on just your side of the conversation, but only messages you sent can be removed from their recipients. You can’t delete from someone else what they sent you, the feature’s PR manager Kat Chui tells me. And Facebook will keep a private copy of the message for a short while after it’s deleted to make sure it can review if it’s reported for harassment.
To use the unsend feature, tap and hold on a message you sent, then select “Remove.” You’ll get options to “Remove for Everyone” which will retract the message, or “Remove for you,” which replaces the old delete option and leaves the message in the recipient’s inbox. You’ll get a warning that explains “You’ll permanently remove this message for all chat members. They can see that you removed a message and still report it.” If you confirm the removal, a line of text noting “you [or the sender’s name] removed a message” (known as a tombstone) will appear in the thread where the message was. If you want to report a removed message for abuse or another issue, you’ll tap the person’s name, scroll to “Something’s Wrong” and select the proper category such as harassment or that they were pretending to be someone else.
Why the 10-minute limit specifically? “We looked at how the existing delete functionality works. It turns out that when people are deleting messages because it’s a mistake or they sent something they didn’t want to send, it’s under a minute. We decided to extend it to 10, but decided we didn’t need to do more,” Chudnovsky reveals.
He says he’s not sure if Facebook’s security team will now resume removing executive messages. However, he stresses that the Unsend button Facebook is launching “is definitely not the same feature” as what was used on Zuckerberg’s messages. If Facebook wanted to truly respect its users, it would at least insert the tombstone when it erases old messages from executives.
Messenger is also building more unsend functionality. Taking a cue from encrypted messaging app Signal’s customizable per thread expiration date feature, Chudnovsky tells me “hypothetically, if I want all the messages to be deleted after six months, they get purged. This is something that can be set up on a per thread level,” though Facebook is still tinkering with the details. Another option would be for Facebook to extend to all chats the per message expiration date option from its encrypted Secret messages feature.
“It’s one of those things that feels very simple on the surface. And it would be very easy if the servers were built one way or another from the very beginning,” Chudnovsky concludes. “But it’s one of those things philosophically and technologically that once you get to the scale of 1.3 billion people using it, changing from one model to another is way more complicated.” Hopefully in the future, Facebook won’t give its executives extrajudicial ways to manipulate communications… or at least not until it’s sorted out the consequences of giving the public the same power.
Truphone — a UK startup that provides global mobile voice and data services by way of an eSIM model for phones, tablets and IoT devices — said that it has raised another £18 million ($ 23.7 million) in funding; plus it said it has secured £36 million ($ 47 million) more “on a conditional basis” to expand its business after signing “a number of high-value deals.”
It doesn’t specify which deals these are, but Truphone was an early partner of Apple’s to provide eSIM-based connectivity to the iPad — that is, a way to access a mobile carrier without having to swap in a physical SIM card, which has up to now been the standard for GMSA-based networks. Truphone is expanding on this by offering a service for new iPhone XS and XR models, taking advantage of the dual SIM capability in these devicews. Truphone says that strategic partners of the company include Apple (“which chose Truphone as the only carrier to offer global data, voice and text plans on the iPad and iPhone digital eSIM”); Synopsys, which has integrated Truphone’s eSIM technology into its chipset designs; and Workz Group, a SIM manufacturer, which has a license from Truphone for its GSMA-accredited remote SIM provisioning platform and SIM operating system.
The company said that this funding, which was made by way of a rights issue, values Truphone at £386 million ($ 507 million at today’s rates) post-money. Truphone told TechCrunch that the funding came from Vollin Holdings and Minden Worldwide — two investment firms with ties to Roman Abramovich, the Russian oligarch who also owns the Chelsea football club, among other things — along with unspecified minority shareholders. Collectively, Abramovich-connected entities control more than 80 percent of the company.
We have asked the company for more detail on what the conditions are for the additional £36 million in funding to be released and all it is willing to say is that “it’s KPI-driven and related to the speed of growth in the business.” It’s unclear what the state of the business is at the moment because Truphone has not updated its accounts at Companies House (they are overdue). We have asked about that, too.
For some context, Truphone most recently raised money almost exactly a year ago, when it picked up £255 million also by way of a rights issue, and also from the same two big investors. The large amount that time was partly being raised to retire debt. That deal was done at a valuation of £370 million ($ 491 million at the time of the deal). Going just on sterling values, this is a slight down-round.
Truphone, however, says that business is strong right now:
“The appetite for our technology has been enormous and we are thrilled that our investors have given us the opportunity to accelerate and scale these groundbreaking products to market,” said Ralph Steffens, CEO, Truphone, in a statement. “We recognised early on that the more integrated the supply chain, the smoother the customer experience. That recognition paid off—not just for our customers, but for our business. Because we have this capability, we can move at a speed and proficiency that has never before seen in our industry. This investment is particularly important because it is testament not just to our investors’ confidence in our ambitions, but pride in our accomplishments and enthusiasm to see more of what we can do.”
Truphone is one of a handful of providers that is working with Apple to provide plans for the digital eSIM by way of the MyTruphone app. Essentially this will give users an option for international data plans while travelling — Truphone’s network covers 80 countries — without having to swap out the SIMs for their home networks.
The eSIM technology is bigger than the iPhone itself, of course: some believe it could be the future of how we connect on mobile networks. On phones and tablets, it does away with users ordering, and inserting or swapping small, fiddly chips into their devices (that ironically is also one reason that carriers have been resistant to eSIMs traditionally: it makes it much easier for their customers to churn away). And in IoT networks where you might have thousands of connected, unmanned devices, this becomes one way of scaling those networks.
“eSIM technology is the next big thing in telecommunications and the impact will be felt by everyone involved, from consumers to chipset manufacturers and all those in-between,” said Steve Alder, chief business development officer at Truphone. “We’re one of only a handful of network operators that work with the iPhone digital eSIM. Choosing Truphone means that your new iPhone works across the world—just as it was intended.” Of note, Alder was the person who brokered the first iPhone carrier deal in the UK, when he was with O2.
However, one thing to consider when sizing up the eSIM market is that rollout has been slow so far: there are around 10 countries where there are carriers that support eSIM for handsets. Combining that with machine-to-machine deployments, the market is projected to be worth $ 254 million this year. However, forecasts put that the market size at $ 978 million by 2023, possibly pushed along by hardware companies like Apple making it an increasingly central part of the proposition, initially as a complement to a “home carrier.”
Truphone has not released numbers detailing how many devices are using its eSIM services at the moment — either among enterprises or consumers — but it has said that customers include more than 3,500 multinational enterprises in 196 countries. We have asked for more detail and will update this post as we learn more.
Instagram users were missing 70 percent of all posts and 50 percent of their friends’ posts before the app ditched the reverse chronological feed for an algorithm in July 2016. Despite backlash about confusing ordering, Instagram now says relevancy sorting has led to its 800 million-plus users seeing 90 percent of their friends’ posts and spending more time on the app.
Yet Instagram has never explained exactly how the algorithm chooses what to show you until today. The Facebook-owned company assembled a group of reporters at its under-construction new San Francisco office to take the lid off the Instagram feed ranking algorithm.
Instagram’s feed ranking criteria
Instagram relies on machine learning based on your past behavior to create a unique feed for everyone. Even if you follow the exact same accounts as someone else, you’ll get a personalized feed based on how you interact with those accounts.
Three main factors determine what you see in your Instagram feed:
- Interest: How much Instagram predicts you’ll care about a post, with higher ranking for what matters to you, determined by past behavior on similar content and potentially machine vision analyzing the actual content of the post.
- Recency: How recently the post was shared, with prioritization for timely posts over weeks-old ones.
- Relationship: How close you are to the person who shared it, with higher ranking for people you’ve interacted with a lot in the past on Instagram, such as by commenting on their posts or being tagged together in photos.
Beyond those core factors, three additional signals that influence rankings are:
- Frequency: How often you open Instagram, as it will try to show you the best posts since your last visit.
- Following: If you follow a lot of people, Instagram will be picking from a wider breadth of authors so you might see less of any specific person.
- Usage: How long you spend on Instagram determines if you’re just seeing the best posts during short sessions, or it’s digging deeper into its catalog if you spend more total time browsing.
Instagram’s team also responded to many of the most common questions and conspiracy theories about how its feed works. TechCrunch can’t verify the accuracy of these claims, but this is what Instagram’s team told us:
- Instagram is not at this time considering an option to see the old reverse chronological feed because it doesn’t want to add more complexity (users might forget what feed they’re set to), but it is listening to users who dislike the algorithm.
- Instagram does not hide posts in the feed, and you’ll see everything posted by everyone you follow if you keep scrolling.
- Feed ranking does not favor the photo or video format universally, but people’s feeds are tuned based on what kind of content they engage with, so if you never stop to watch videos you might see fewer of them.
- Instagram’s feed doesn’t favor users who use Stories, Live, or other special features of the app.
- Instagram doesn’t downrank users for posting too frequently or for other specific behaviors, but it might swap in other content in between someone’s if they rapid-fire separate posts.
- Instagram doesn’t give extra feed presence to personal accounts or business accounts, so switching won’t help your reach.
- Shadowbanning is not a real thing, and Instagram says it doesn’t hide people’s content for posting too many hashtags or taking other actions.
Today’s Instagram whiteboard session with reporters, its first, should go a long way to clearing up misunderstandings about how it works. When people feel confident that their posts will reach their favorite people, that they can reliably build a public audience, and that they’ll always see great content, they’ll open the app more often.
Yet on the horizon looms a problem similar to what Facebook’s algorithm experienced around 2015: competition reduces reach. As more users and businesses join Instagram and post more often, but feed browsing time stays stable per user, the average post will get drowned out and receive fewer views. People will inevitably complain that Instagram is trying to force them to buy ads, but it’s a natural and inevitable consequence of increasingly popular algorithmic feeds.
The more Instagram can disarm that problem by pushing excess content creation to Stories and educating users about how the feed operates, the less they’ll complain. Facebook is already uncool, so Instagram must stay in our good graces.
The science of prosthetics has been advancing by leaps and bounds over the last few years, and research into soft robotics has been especially complementary. The same techniques that go into making a robot arm that flexes and turns like a real one can go into making more complex, subtle organs — like the heart, as Swiss researchers have demonstrated. Read More