Start Up No.1,125: Chrome’s unused extensions, Alexa, Google and Siri stop human listeners, scooters v the climate, can the internet change abortion?, and more

Google wants companies to bid to be its default search on mobile. Does that seem reasonable? CC-licensed photo by Paulo O on Flickr.

You can sign up to receive each day’s Start Up post by email. You’ll need to click a confirmation link, so no spam.

A selection of 11 links for you. Not part of an auction. I’m @charlesarthur on Twitter. Observations and links welcome.

Half of all Google Chrome extensions have fewer than 16 installs • ZDNet

Catalin Cimpanu:


There are 188,620 extensions available on the Chrome Web Store, and while you might think this provides a wide variety of choices for Chrome users, in reality, most of these extensions are dead or dwindling, with very few having active installations.

All in all, about 50% of all Chrome extensions have fewer than 16 installs, meaning that half of the Chrome extension ecosystem is actually more of a ghost town, according to a recent scan of the entire Chrome Web Store conducted by Extension Monitor.

Further, 19,379 extensions (just over 10%) have zero installs, and 25,540 extensions (13% of the total) have just one user.

The scan found that there are very few Chrome extensions that managed to establish a dedicated userbase.

According to Extension Monitor, around 87% of all extensions have fewer than 1,000 installs, a number that many extension devs would consider a failure, taking into account that the Chrome browser has over one billion monthly active users, a huge potential market for any extension developer.

At the other side of the spectrum, only 13 extensions have managed to break over the 10 million mark — the highest user count threshold available on the Chrome Web Store.

Those 13 are Google Translate, Adobe Acrobat, Tampermonkey, Avast Online Security, Adblock Plus, Adblock, uBlock Origin, Pinterest Save Button, Cisco Webex, Grammarly for Chrome, Skype, Avast SafePrice, and Honey.


23.5% have one or zero installs – so there’s another 50% with between one and 1,000 installs. It’s very heavily weighted to the “nobody uses this” end, calling into question their whole existence.
unique link to this extract

Google to ask rivals to bid to be default search on Android phones • Bloomberg

Natalia Drozdiak:


Alphabet’s Google will require rivals to bid in order to become listed as alternative search providers on Android smartphones, a move to try to keep additional antitrust scrutiny at bay.

Starting next year, Google will prompt users to make a choice between Google and three other rival options as their default search provider. Google invited search providers to bid as part of an auction on the new choice screen, which will appear when a user sets up a new Android smartphone or tablet in Europe for the first time.

The European Commission, the bloc’s antitrust body, last year fined Google €4.3bn ($4.8bn) for strong-arming device makers into pre-installing its Google search and Chrome browser, giving it a leg up because users are unlikely to look for alternatives if a default is already preloaded. The EU ordered Google to change that behavior and threatened additional fines if it failed to comply.

Eric Leandri, chief executive of Paris-based search engine Qwant, called Google’s move “a total abuse of the dominant position” to “ask for cash just for showing a proposal of alternatives.”

…A European Commission spokeswoman said the EU would be “closely monitoring the implementation of the choice screen mechanism” and noted that the changes allow rival search engines the possibility to strike deals with smartphone and tablet manufacturers to pre-install their services.


Seems fair, as long as Google is obliged to bid, and its losing bid price goes to the winner; or if Google has the highest bid, the money is distributed to the other bidders. (Or I’m sure you can think of a better distribution system.) After all, if the EU says Google got its dominant position through monopoly abuse, why should it be allowed to continue monetising it?
unique link to this extract

Sorry, scooters aren’t so climate-friendly after all • MIT Technology Review

James Temple:


the mere fact that battery-powered scooters don’t belch pollution out of a tailpipe doesn’t mean they’re “emissions free,” or as “eco-friendly” as some have assumed. The actual climate impact of the vehicles depends heavily on how they’re made, what they’re replacing, and how long they last.

Researchers at North Carolina State University decided to conduct a “life-cycle assessment” that tallied up the emissions from making, shipping, charging, collecting, and disposing of scooters after one of them noticed that a Lime receipt stated, “Your ride was carbon free.”

The study concludes that dockless scooters generally produce more greenhouse-gas emissions per passenger mile than a standard diesel bus with high ridership, an electric moped, an electric bicycle, a bicycle—or, of course, a walk.

The paper found that scooters do produce about half the emissions of a standard automobile, at around 200 grams of carbon dioxide per mile compared with nearly 415. But, crucially, the researchers found in a survey of e-scooter riders in Raleigh, North Carolina, that only 34% would have otherwise used a personal car or ride-sharing service. Nearly half would have biked or walked, 11% would have taken the bus, and 7% would have simply skipped the trip.

The bottom line: roughly two-thirds of the time, scooter rides generate more greenhouse-gas emissions than the alternative. And those increased emissions were greater than the gains from the car rides not taken, says Jeremiah Johnson, an engineering professor and one of the authors of the paper.


Individual devices are less fuel-efficient than collective ones. Though the results for the electric bicycle are surprising. In general, most of the life cycle costs are in the materials.
unique link to this extract

Abortion pills should be everywhere • The New York Times

Farhad Manjoo (who’s a bloke):


most of my orders came through fine. Each of the three pill packages I got cost me between $200 and $300, including expedited shipping. (The average cost of an abortion in the United States is about $500.)

I spent months looking for a lab that would test my pills; many waved me off, wary of controversy. Finally, I got in touch with Alan Wu, chief of the clinical chemistry laboratory at San Francisco General Hospital, whose lab tested a couple of my mifepristone tablets. The finding: They were authentic. I wasn’t surprised; in a more comprehensive study conducted by Gynuity Health and Plan C, published last year in the journal Contraception, researchers in four states ordered abortion pills from 16 different online pharmacies, and found they were all just what they said they were.

Each time I got a pack of pills in the mail, I was increasingly bowled over: If this is so easy, how will they ever stop this? I’ve been watching digital markets for 20 years, and I’ve learned to spot a simple, powerful dynamic: When something that is difficult to get offline becomes easy to get online, big changes are afoot…

…The activists building the online pill network acknowledge that there are potential dangers in the market — but they insist that the risks are far smaller than many guess.


Activists reckon it could lead to more informal early-stage abortions, which is more politically acceptable (and would be a lot simpler to do in states with absurdly early abortion limits). Would this be the first political change driven solely by availability of treatment through the internet?
unique link to this extract

Richard Thaler: ‘If you want people to do something, make it easy’ • Financial Times

Tim Harford:


The key message of [Thaler’s book] Nudge was that governments could improve the health and wellbeing of their citizens without infringing on their liberty, simply by more thoughtfully designing their rules, procedures, or even labelling.

“If you want people to do something, make it easy.” Put the cashews in the kitchen and the fruit by the cafeteria checkout.

More recently, Thaler has been thinking and writing about what he calls “sludge”. It’s the same procedure in reverse: if you want people not to do something, make it difficult.

Reaching for an example, Thaler has a bone to pick with The Times. The first review of Misbehaving was published there, and Thaler’s editor sent him a link.

“And I can’t get past the paywall without subscribing.”

But then he notices there’s an offer of a month’s trial subscription at an introductory rate. “But I read further, having written a book about this, and I see that it will be automatically renewed.”

Not only that, it will be renewed at full price, “and that in order to quit, I have to give them 14 days’ notice. So the one month free trial is actually two weeks. And I have to call London [from Chicago] in London business hours, not on a toll free line.”

He pauses and chides me to check that the FT isn’t placing similar sludge in the way of readers who wish to unsubscribe. I assure him that nobody would ever want to unsubscribe, but in any case such knavery would be beneath us. But part of me wonders. “Check your policy at the FT,” he advises.


“Sludge” is a neat idea – in web design you’d probably call it dark patterns. There’s plenty more, particularly about Brexit pronouncements and about the announcement to “mind the gap” on the London Underground.
unique link to this extract

Tink Labs set to shut down amid mass layoffs • Financial Times

Siddarth Shrikanth and Mercedes Ruehl:


Tink Labs, which was founded in 2012, was one of Hong Kong’s best funded startups. Investors include Foxconn subsidiary FIH Mobile; Cai Wensheng, chairman of popular Chinese selfie app Meitu; and Sinovation Ventures, an investment fund headed by former Google China chief Kaifu Lee. SoftBank’s mobile unit invested via a joint venture with Tink in Japan.

According to several current and former employees, Tink Labs has said it will close on Thursday, after mass layoffs in recent weeks. The company did not respond to requests for comment.

At its height, Tink Labs was valued at as much as $1.5bn, and its “Handy” smartphones service had handsets in more than 600,000 hotel rooms across 82 countries, via relationships with big hotel chains including Hyatt Hotels, InterContinental Hotel Group and Shangri-La Hotels and Resorts.

The closure will see Tink Labs join a lengthening list of Chinese startups that have collapsed.

Bicycle-sharing company ofo went from world-leading “sharing economy” startup and tech darling to the verge of bankruptcy in just four years. Rival Bluegogo has folded, while Aiwujiwu, a Chinese online property listings platform backed by Hillhouse and Temasek, reportedly went into liquidation earlier this year.

Meanwhile, the flow of capital into China’s tech sector has begun drying up, while due diligence on prospective investments has increased significantly as investors grow wiser to potential risks.


unique link to this extract

Amazon gives option to disable human review of Alexa recordings • Bloomberg

Matt Day:

» Inc. will let Alexa users opt out of human review of their voice recordings, a move that follows criticism that the program violated customers’ privacy.

A new policy took effect Friday that allows customers, through an option in the settings menu of the Alexa smartphone app, to remove their recordings from a pool that could be analyzed by Amazon employees and contract workers, a spokeswoman for the Seattle company said. It follows similar moves by Apple Inc. and Google.

Bloomberg first reported in April that Amazon had a team of thousands of workers around the world listening to Alexa audio requests with the goal of improving the software. Their tasks include listening to and transcribing voice recordings. Some of the workers reviewing customer recordings had access to certain personal data, including users’ first names and their location.


Yeah, but nobody took any notice of Bloomberg’s report in April, because it wasn’t written in a way that grabbed people. Now, here we are a week after an explosive Guardian report, and all three organisations have, for one reason or another, turned off human review. Perhaps they all proceeded down their own timelines to get to the same place at the same time; that implies that Apple’s a lot quicker to get there, Google next fastest, and Amazon a bit tardy.
unique link to this extract

Google ordered to halt human review of voice AI recordings over privacy risks • TechCrunch

Natasha Lomas:


A German privacy watchdog has ordered Google to cease manual reviews of audio snippets generated by its voice AI. 

This follows a leak last month of scores of audio snippets from the Google Assistant service. A contractor working as a Dutch language reviewer handed more than 1,000 recordings to the Belgian news site VRT which was then able to identify some of the people in the clips. It reported being able to hear people’s addresses, discussion of medical conditions, and recordings of a woman in distress.

The Hamburg data protection authority told Google of its intention to use Article 66 powers of the General Data Protection Regulation (GDPR) to begin an “urgency procedure” under Article 66 of GDPR last month.


Surprise: Google complied. It told Ars Technica that “Shortly after we learned about the leaking of confidential Dutch audio data, we paused language reviews of the Assistant to investigate. This paused reviews globally.” No date for resumption.
unique link to this extract

Apple halts practice of contractors listening in to users on Siri • The Guardian

Alex Hern:


Contractors working for Apple in Ireland said they were not told about the decision when they arrived for work on Friday morning, but were sent home for the weekend after being told the system they used for the grading “was not working” globally. Only managers were asked to stay on site, the contractors said, adding that they had not been told what the suspension means for their future employment.

The suspension was prompted by a report in the Guardian last week that revealed the company’s contractors “regularly” hear confidential and private information while carrying out the grading process, including in-progress drug deals, medical details and people having sex.

The bulk of that confidential information was recorded through accidental triggers of the Siri digital assistant, a whistleblower told the Guardian. The Apple Watch was particularly susceptible to such accidental triggers, they said. “The regularity of accidental triggers on the watch is incredibly high … The watch can record some snippets that will be 30 seconds – not that long, but you can gather a good idea of what’s going on.


One week from the original report to this change. That’s impressive – moreso given that Bloomberg had a weaker form of this report much earlier this year but didn’t get anything like the detail. The power of newsprint: it makes a difference having something you can put on a chief executive’s desk (even if you have to fly it out there).

Apple has indicated that it’s eventually going to restart this, but on an opt-in basis.
unique link to this extract

The new blood test for Alzheimer’s disease: developed in a study without patients • Medium

Cecile Janssens:


It is these opening lines of the study’s press release that shaped the news. The Alzheimer’s Society reports that “Blood test is 94% accurate at identifying early Alzheimer’s disease”; The Guardian that “Alzheimer’s blood test could predict onset up to 20 years in advance”; and also the doctors at WebMD highlight that “Blood test may spot signs of early Alzheimer’s.”

But no, the study didn’t test and track people for 20 years to see who ultimately developed Alzheimer’s disease. And the test wasn’t 94% accurate in identifying early Alzheimer’s either. None of the participants in the study was diagnosed with Alzheimer’s disease. Their average score on the Mini Mental State Examination, a well-known test to measure cognitive impairment, was 29. As a reference: the test’s best possible score is 30, a score of 20 to 24 may indicate mild dementia, and lower than 12 severe dementia.

The study wasn’t about prediction either. The claim that indications of brain amyloidosis can be observed two decades before the first symptoms appear must have come from other studies. I didn’t find citations to these studies in the article. (I wonder whether such prediction studies exist.)


Basically, it’s not at all what it appears to be, which is disappointing, but – in my experience – completely ordinary for reports about medical studies (and even press releases about medical studies).
unique link to this extract

Do smartphones need gesture HMI? • Strategy Analytics

Paul Brown on the promised “gesture” control for the forthcoming Google Pixel 4:


Gestures are not something new to smartphones. In 2013, Samsung introduced the Galaxy S4 with a host of gestures. However, most of these gestures were cumbersome and inefficient, had low adoption, and many were removed from future Samsung devices.

According to Google’s blog post, the number of initial gestures on the Pixel 4 will allow the user to undertake the following three functions, just by waving your hand:

• Skip songs
• Snooze alarms
• Silence phone calls

Using gestures to snooze alarms and silence phone calls could be very useful.  These are both tasks that will likely occur when the user is not holding the phone. Waving a hand over the phone when either event occurs is a very simple action, and one that requires less cognitive effort than picking up the phone and pressing buttons (physical or on the touchscreen). However, there may be a concern that the user accidentally silences a phone call when they move their hand towards the phone to pick it up and answer the call. The required gesture and how it can differentiate a user’s intent is key here.


Samsung’s S4’s “Air Gestures” were amazingly annoying. As Brown points out, with the Pixel, if the gesture doesn’t work when the display isn’t lit (eg to skip the song), then you’ll need to tap it to then gesture. In which case you might as well wake-and-tap. But if it works when the display is off, the potential for accidental gesturing is huge. I’m not convinced.
unique link to this extract

Errata, corrigenda and ai no corrida: none notified

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.