Monthly Archive March 2018


Some hard truths about Twitter’s health crisis

It’s a testament to quite how control freaky and hermetically sealed to criticism the tech industry is that Twitter’s CEO Jack Dorsey went unscripted in front of his own brand livestreaming service this week, inviting users to lob awkward questions at him for the first time ever.

It’s also a testament to how much trouble social media is in. As I’ve written before, ‘fake news’ is an existential crisis for platforms whose business model requires them to fence vast quantities of unverified content uploaded by, at best, poorly verified users.

No content, no dice, as it were. But things get a whole lot more complicated when you have to consider what the content actually is; who wrote it; whether it’s genuine or not; and what its messaging might be doing to your users, to others and to society at large.

As a major MIT study looking at a decade’s worth of tweets — and also published this week — underlines: Information does not spread equally.

More specifically, fact-checked information that has been rated true seems to be less sharable than fact-checked information that has been rated false. Or to put it more plainly: Novel/outrageous content is more viral.

This is entirely unsurprising. As Jonathan Swift put it all the way back in the 1700s: “Falsehood flies, and the Truth comes limping after it.” New research, old truth.

What’s also true is that as social media’s major platforms have scaled, so too have the problems blasted through their megaphones zoomed into mainstream view.

Concerns have ballooned. We’re now at a structural level, debating societal fundamentals like cohesion, civility, democracy. Even, you could argue, confronting humanity itself. Platform as a term has always had a dehumanizing ring. Perhaps that’s their underlying truth too.

Dorsey says the “health” of conversations on his platform is now the company’s “number one priority” — more than a decade after he typed that vapid first tweet, “just setting up my twttr”, when he presumably had zero idea of all the horrible things humans would end up using his technology for.

But it’s also at least half a decade after warnings that trolls and bots were running rampant on Twitter’s platform.

Turns out the future comes at you eventually. Even if you stubbornly refuse to listen as alarm after alarm are being sounded. “Never send to know for whom the bell tolls; it tolls for thee,” wrote John Donne, meditating on society and the individual, back in 1624.

#280 assessment of what a buzzcut, bearded and careworn Dorsey now says he sees as Twitter’s main problem and thus priority boils down to something like this…

We know our platform is being used negatively, people are hurting and public conversation is being damaged. But we don’t know how to fix it because we don’t understand how to measure the individual and societal impacts of our technology. We think more tech can help. Pls help us.

What Twitter’s crisis tells us is that tech companies are terrible listeners. Although those of us outside the engineering room knew that already.

It’s hardly a surprise that techies suck at listening when they sit inside their hermetically sealed pods thinking it’s both their special gift and libertarian right to control levers that remotely affect other people’s lives while channelling the spice and dollars their way.

So it is a good sign, albeit horribly overdue, to see a nervous and contrite-seeming Dorsey stand in front of the firehose of user opinion — for 50 or so raw, unedited minutes.

Hopefully this performance — which he said would be repeated regularly, from here on in — signals an absolute conversion to reformation. A realization that social media platforms can’t engineer around societal responsibility. That listening and understanding is absolutely their day job.

Head-in-the-sand-ism will catch up with you eventually. Just as playing fast and loose finally overtook Uber’s founder and landed his company in all sorts of legal hot water.

So how did Dorsey and select members of his safety ‘A-team’ do in their first ‘awkward questions’ Periscope?

Fair to middling, is my assessment. It’s clear they still don’t really know how to fix the mess they are in. Hence Twitter soliciting proposals from the public. But admitting they don’t know what to do and reaching out for help is a big and important step.

To put it colloquially, they’ve realized the shit they’re in. And the shit that’s at stake. Hashtag #changeforreal

Dorsey seemed visibly uncomfortable with the Periscope process, which again is testament to how closed a box and operating shop Twitter has been. He hasn’t always been CEO but he is a founder so he’s absolutely on the hook for that.

And Twitter’s bunker mentality has clearly compounded its problems in identifying and responding to content issues that first flared on its platform and then raged. Unpicking that won’t be easy.

Indeed, he said several times that the changes he wants to happen “won’t happen overnight”. That changing Twitter will require a lot of work.

He also admitted the company has “a lot of historical divisions” and said it has not always been as collaborative as it could have. tl;dr inside Twitter there’s a bunch of other bunkers — which truly sounds like a culture nightmare.

So when he talked about the hard work coming I don’t think Dorsey just meant reengineering lots of systems and cranking out lots more user surveys. Because changing an ingrained culture and its processes is a beast. Which is why it’s much better to start from a place of enlightenment. But hey, silver lining, here Twitter finally, finally is, admitting it screwed up and wanting to start over.

At least it’s now saying it wants its product to have a holistic and healthy impact on the world. That it wants to try and reset the coarsening of public discourse that social media has wrought. Certainly it’s a more evolved mission statement than its previous one — which was basically: ‘Eat our free speech.’

That said, Dorsey’s focus on a new type of measurement — this idea of a ‘health metric’ — as the solution for toxic content seems to me problematic. Almost, you could say, like the trigger response of an engineer confronting an ethics textbook for the first time.

Because Twitter’s content problems really boil down to Twitter failing to enforce the community standards it already has. Which in turn is a failure of leadership, as I have previously argued.

A good current example is that it has an ads policy that bans “misleading and deceptive” ads. Yet it continues to accept advertising money from unregulated entities pushing dubiously obscure crypto exchanges and flogging wildly risky token sales.

Twitter really doesn’t need to wait for a new metric to understand that the right thing to do here is to take crypto/ICO ads off its platform right now.

Shucks, even Facebook has done this.

Yet Dorsey and his team omitted to mention ads when he was asked about crypto scams during the Periscope. They just talked about what they’re doing to tackle Twitter users trying to tweet-scam others into sending a bit of crypto.

Continuing to accept ad money attached to what’s still an essentially unregulated space, when there are so many visible and public concerns because scams really are part of the furniture, really is indefensible. Banning these ads is both common sense and just the right thing to do.

And so if Twitter needs to wait for someone else to invent some kind of holistic wellness metric in order to make that low-hanging Satoshi drop then, well, its culture change is going to be much harder and much more painful than Dorsey imagines.

Obsession with measurement and the search for a universal problem-solving metric — to try to quantify the “health, openness and civility of public conversation”, as Twitter puts it — also looks very much like a strategy to buy time.

It may ultimately turn out to be misdirection too; an attempt to deflect blame and divert criticism via solutioneering.

By outsourcing a challenge, and seeking to co-opt the energy and ideas of third parties, Twitter is also reframing what’s broken in a way that starts to spread responsibility for the problems its platform is causing. (Maybe it’s taken a leaf out of Facebook’s playbook on that.)

Content moderation is certainly a hard problem if you understaff it. But if you employ enough machine-aided humans to properly enforce your community standards then it’s quite possible to shrink a toxic content problem.

Throw enough resources in and content problems can become vanishingly small, even insignificant. This is known as community management.

Yes there are counter risks. Especially if, like Twitter, you’ve historically advertised yourself as the free speech wing of the free speech party.

But if you’re having trouble drawing service red lines around, for example, known neo nazis, for whom hate speech and agitating for violence is a way of life, then setting out on a long and winding quest to deconstruct the anatomy of society in the hopes of eventually being able to build algorithms that do a better job of keeping toxic content off your platform, well, that probably isn’t the fundamental fix you should be searching for.

The problem right now is that Twitter doesn’t have the courage — or, heck, the imagination — to enforce its own community guidelines.

Though the hard truth may well be that it just cannot afford to. That the business model never did stack up. Not if you have to factor in the cost of staffing up to properly moderate all the shit that’s being uploaded and thrown about.

Meanwhile the costs of toxic, hate inciting messages blitzkrieging public conversation via the amplifying megaphone of social media keep on rising…

In his Periscope plea for help, Dorsey also said he wants Twitter to be “one of the most trusted services in the world”. But if he thinks he can build a for-all-technotopia where liberals co-exist peacefully alongside neo nazis — thanks to a shiny new set of augmented reality controls that fade view from counter view — he’s still thinking fatally inside the tech industry black box.

Social media has always bled offline. Its wounds, like its users, are human. Its shaping impacts are felt by people and across society.

Another old truth: You can’t please all of the people, all of the time. So if Dorsey thinks he can find a technology fix for that age-old challenge he’s going to waste a whole lot more money and a whole lot more time — while the rest of us bleed.




Android beats iOS in smartphone loyalty, study finds

Samsung’s new Galaxy S9 may not quite live up to the iPhone X when it comes to Samsung’s implementation of a Face ID-style system or its odd take on AR emoji. But that’s not going to matter much to Samsung device owners — not only because the S9 is a good smartphone overall — but because Android users just aren’t switching to iPhone, at least not like they used to.. In fact, Android users have higher loyalty than iOS users do, according to a new report today from Consumer Intelligence Research Partners (CIRP).

The research firm found that Android brand loyalty has been remaining steadily high since early 2016, and remains at the highest levels ever seen.

Today, Android has a 91 percent loyalty rate, compared with 86 percent for iOS, measured as the percentage of U.S. customers who stayed with their operating system when they upgraded their phone in 2017.

From January 2016 through December 2017, Android loyalty ranged from 89 to 91 percent (ending at 91 percent), while iOS loyalty was several percentage points lower, ranging from 85 to 88 percent.

Explains Mike Levin, partner and co-founder of CIRP, users have pretty much settled on their brand of choice at this point.

“With only two mobile operating systems at this point, it appears users now pick one, learn it, invest in apps and storage, and stick with it. Now, Apple and Google need to figure out how to
sell products and services to these loyal customer bases,” he said.

That’s also why both companies have increasingly become focused on services, as they try to extract larger revenues from their respective user bases. For Apple, that’s been a win, financially speaking — it saw record revenue from services in November, suggesting growth in things like Apple Music, Apple Pay, iCloud, AppleCare and App Store.

For Android users, the higher brand loyalty could be chalked up to their ability to switch to different styles of new phones, without having to leave Android — thanks to its distribution across a variety of manufacturers’ handsets. That gives users the freedom to try out new experiences, without giving up their investments in purchased apps, or the time they’ve spent learning their way around Android, for that matter.

It’s worth noting that Android hasn’t always led in user loyalty as it does now. CIRP has been tracking these metrics for years, and things used to be the other way around.

In 2013, for example, iPhone owners were found to be more loyal than Android users. But that shifted the following year, and Android has risen ever since. (By the way, if you click through to read the comments on that linked AllThingsD article from 2013, it’s a quite a trip. Remember when people cared so much about their choice of smartphone it led to commenting wars? Ah, the good ol’ days.)

All that being said, the rate of switching is different from the total number of people switching, the firm also pointed out. And looking at the numbers from that perspective changes things.

“We know Android has a larger base of users than iOS, and because of that larger base, the
absolute number of users that switch to iOS from Android is as large or larger than the
absolute number of users that switch to Android from iOS,” said Levin.”Looking at absolute number of users in this way tends to support claims that iOS gains more former Android users,
than Android does former iOS users.”



Cellebrite may have found a way to unlock iPhones running iOS 11

According to a Forbes report, Israeli company Cellebrite is now able to unlock some very recent iPhones. Cellebrite is a well-known company that sells mobile forensics tools to extract data from locked devices.

While early versions of iOS weren’t really secure, this has changed quite a lot in recent years. All iOS devices now ship with a secure enclave, all data is encrypted if you use a passcode and there are multiple security checks when you boot and use your device.

In other words, if you don’t have the passcode, you’re going to have a hard time getting your hand on the data on the device. Many firms try to find vulnerabilities to unlock mobile devices. It has become a lucrative industry as intelligence agencies often pay forensics companies to unlock mobile devices.

Those forensics methods often lag behind. For instance, it’s quite easy to find a device to unlock an iPhone 6 running iOS 8. But if Forbes’ report and Cellebrite’s website are right, governments can now pay Cellebrite to unlock an iPhone 8 running iOS 11. It’s also worth noting that Cellebrite can unlock recent Android devices as well.

It’s unclear if it works with the most recent version of iOS 11 (11.2.6) or just the operating system version that was available back in September (11.0). It’s also unclear if it works with all iOS devices or if it only works with some devices. Forbes found a warrant that mentions an unlocked iPhone X.

This is a cat-and-mouse game, and Apple engineers are now probably working hard to fix all the vulnerabilities they can find. As always, if you don’t want to let authorities read your personal data, you should keep your devices up-to-date.

In addition to new features, security patches protect you against the most common attacks. And malicious hackers might use the same vulnerabilities against you.




FCC accuses stealth space startup of unauthorized satellite deployment

The FCC has denied a space startup permission to launch a collection of communications satellites after discovering that it had already launched some — after being told not to. Swarm Technologies, still in stealth mode, appears to have gone ahead with the deployment of four satellites deemed too small to be tracked and therefore unsafe to put into orbit.

IEEE Spectrum put the pieces together from public FCC documents and some launch manifests. Swarm’s original plan was to put several very small satellites — smaller even than 1U Cubesats — in orbit to test its experimental communications system.

But the small size meant the satellites couldn’t be tracked with existing space monitoring technology, and the FCC, which must approve communications satellite launches, considered this too great a risk and declined to authorize Swarm’s proposed deployment.

What should have happened next is: Swarm scrubs the deployment, applies again with larger satellites or some other means of improving the small ones’ visibility, the FCC grants permission and then the launch happens.

While the company did reapply with larger satellites, it seems to have gone ahead with the original plan of launching the tiny satellites despite the FCC’s warning not to. This is evident from the manifest of India’s Polar Satellite Launch Vehicle (PDF) that took off in January, which included four “SpaceBEEs” matching the description of Swarm’s unauthorized craft.

It’s possible that Swarm’s satellites were already locked and loaded, and perhaps more importantly, paid for, by the time the FCC issued their decision in December. The long lead times for both approval and launch mean that much prep must be done while a deployment is still waiting for the official go-ahead — if you waited for the red tape to clear before even applying for a launch spot, you might run out of funding just waiting for your chance to get into orbit.

But in this case, especially as the FCC cited a safety issue — the inability to reliably track the satellites’ location — the correct thing to do would be to pull out of the launch. That’s easy for me to say, of course, it’s not my money or company, but skirting the rules like this may prove more costly in the end than adhering to them.

I’ve asked Swarm, the FCC and Spaceflight (which appears to have arranged Swarm’s spaceon the launch, perhaps thinking authorization was forthcoming) for comment and will update this story if I hear back.




Giphy held talks to raise a massive new funding round

We’re hearing from a number of sources that Giphy, the big platform for hosting GIFs that also runs a GIF keyboard, held talks to raise a huge new financing round — though it’s not clear if it ever crossed the finish line.

Sources pegged the round at something as high as around $100 million, but that may have changed over time. We’ve been hearing about this for some time now, and whispers of this seem to have started a few months ago. As always, it’s possible that the talks may have changed over time. In the end, Giphy may have not have gone with financing at all, and for the time being, we’re hearing that nothing is imminent. Giphy last raised $72 million at a reported $600 million valuation at the end of 2016.

But the consumer investing environment isn’t necessarily dead, or even in purgatory, right now. HQ Trivia, for example, was able to raise $15 million at a $100 million valuation. This comes amid a time when the GIF space at large seems to be heating up. Given that the space seems to be growing quickly, it makes sense to try to raise additional capital in order to secure the right partnerships — and also get the right talent on board to optimize the experience so users are getting the right GIFs at the right moments and keep coming back to the platform over and over again. Given the growth, and that the business model isn’t fully fleshed out, it makes sense that Giphy could use some additional cash.

The apps in the space clearly have momentum. Giphy says it has 300 million daily active users — which, depending on who you ask in the Valley, could have a number of different interpretations. One of Giphy’s competitors, Tenor, points to searches on its platform as a success metric — saying that it hit 12 billion GIF searches in February. Gfycat, meanwhile, is positioning itself as a company geared around creator tools with mechanisms that optimize the fidelity of the inbound GIF, which also says it has around 130 million monthly active usersGfycat raised $10 million in 2016, while Tenor (formerly Riffsy) has raised around $32 million total.

It also presents a unique opportunity for all these platforms to start thinking about sponsored content. For example, if you open up a GIF search engine inside of a keyboard, one of these companies could plant a sponsored GIF right inside the search rail. Should it be sticky enough and hit the right sweet spot, it could get incredibly high share counts, and as a result offer a lot of reach for those companies looking to make GIFs.

This kind of branded content model is usually tied in with messaging, but GIFs could offer leagues more engagement than the average ad — which is what advertisers are looking for. Tenor offers some sponsored GIF products already, for example.

You’ll find a lot of Gfycat links around the internet, but some of the most fertile ground for these platforms exists within the various messenger platforms. Facebook Messenger, for example, uses these platforms more or less indiscriminately — switching between services relatively easily as it looks to just optimize the user experience and give them the best content. But for iMessage, for example, users install a specific keyboard. Neither of these apps are exactly blockbusters (nor should they be compared to apps like Facebook).

Here’s the Giphy app, where you can search for GIFs and copy them and such, for the last 90 days:

GIFs are increasingly popular, partly thanks to their ability to compress a ton of information into a short clip. This compression allows for punchy, memorable communication, which is great for messaging, but also great for ads.

While you could easily write out a text that tries to translate that information, searching for a GIF that translates not just the text but also the kind of subtext offers a ton of value. It’s thanks to that these platforms have risen to such prominence — both with Giphy’s own 300 million daily active user number and Tenor’s 12 billion monthly searches number. They take different approaches to measuring their success, but the point remains that this represents a pretty massive opportunity.

We reached out to Giphy several times for comment, but did not hear back. We’ll update the story when we hear back from them.




Intelligo is using AI to make background checks relevant again

To realize that the background check industry needs an overhaul look no further than the backlog of 700,000 background checks faced by the federal agency that handles all background checks for sensitive government positions. This backlog has essentially rendered background checks useless, as many agencies are able to give security clearances on a temporary basis before a background check is even started.

Intelligo is an Israeli company trying to make background checks relevant again by using AI and machine learning to not only speed up and automate the process, but also run more thorough checks.

Launching out of beta today, the company has raised $6.8M to date – a seed round of $1.1M and a Series A of $5.7M. They boast investors like Eileen Murray (Co-CEO of Bridgewater Associates) and advisors like the former director of the NSA Michael McConnell and former Managing Director of the Israel Ministry of Defense Pinhas Buchris.

Currently most serious background checks are done manually. This means that when an analyst creating a report comes across a new data source they need to decide if it’s worth taking the time to parse it and add it to the report. Consequently, many important sources like social media pages and news sites are left out of reports. It also means that background checks can take up to a week or longer, which is frustrating for the company and applicant.

Alternatively, Intelligo’s solution is primarily driven by an automated machine learning platform that can indiscriminately look at all thousands of data sources without concern for how much manual labor it will take. Reports are also provided in a user-friendly interactive dashboard, which is a stark contrast to the dozens of typed pages that an old-school background check will be.

Automating the process also dramatically cuts down on cost – Intelligo says their prices are half of the average market price, which is allowing small and midsize businesses to now get the benefit of a high-level background check that typically would only be used by a larger corporation.

The startup also offers an ongoing monitoring product designed for the investment world. Funds often want the ability to monitor their portfolio companies and management teams even after the initial due diligence process, and by using an automated platform Intelligo can let funds know of management issues long before a human would find the source of the issue.