Start Up No.1,139: how bots can change votes, the drone bust, pricing the Galaxy Fold, Sonos gets mobile, and more

YouTube knew underage kids were watching videos on its site, and was fined; now content creators will pay the price. CC-licensed photo by Jon Pinder 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 10 links for you. Friday already? I’m @charlesarthur on Twitter. Observations and links welcome.

How social networks can be used to bias votes • Nature

Nature Editorial Board:


Politicians’ efforts to gerrymander — redraw electoral-constituency boundaries to favour one party — often hit the news. But, as a paper published in Nature this week shows, gerrymandering comes in other forms, too.

The work reveals how connections in a social network can also be gerrymandered — or manipulated — in such a way that a small number of strategically placed bots can influence a larger majority to change its mind, especially if the larger group is undecided about its voting intentions (A. J. Stewart et al. Nature 573, 117–118; 2019: “Information gerrymandering and undemocratic decisions”).

The researchers, led by mathematical biologist Alexander Stewart of the University of Houston, Texas, have joined those who are showing how it can be possible to give one party a disproportionate influence in a vote.

It is a finding that should concern us all.


From the paper:


Our mathematical analysis uncovers a phenomenon that we call information gerrymandering: the structure of the influence network can sway the vote outcome towards one party, even when both parties have equal sizes and each player has the same influence. A small number of zealots, when strategically placed on the influence network, can also induce information gerrymandering and thereby bias vote outcomes. We confirm the predicted effects of information gerrymandering in social network experiments with n = 2,520 human subjects.


unique link to this extract

Lenovo Mirage AR headset with Marvel games goes on sale for $250 • Variety

Janko Roettgers:


called the Lenovo Mirage AR headset, the device once again relies on a consumer’s phone, and an app that can be downloaded for free, to super-impose pictures over their view of the real world. “You are still grounded in your world,” said Lenovo senior product marketing manager Wahid Razali. “You are bringing the heroes into your space.”

And while the first iteration of the headset shipped with lightsaber controllers, this new version comes with a pair of more generic grip controllers that can be used to power a variety of games.

When Lenovo came out with the first iteration of the headset, the two companies tried a variety of games, including their own take on holochess. Turns out that players care a lot more about fighting Stormtroopers than playing chess in AR, which is why the two companies refocused on life-sized battles for their new collaboration.

In the case of “Star Wars: Jedi Challenges,” the game allows players to turn into Doctor Strange, Captain America, Thor, Black Panther, Captain Marvel and Star-Lord, and face off against adversaries like Loki and the Winter Soldier. “You’ll be playing as iconic heroes fighting iconic villains,” said Razali.

In addition to a story mode that allows those one-on-one face-offs, the game also supports a survival mode that tasks players with fighting back waves of enemies, and a co-op mode that lets multiple players team up, and compete for the highest score. The latter naturally requires multiple headsets, which won’t come cheap: At launch, the new Lenovo Mirage AR headset retails for $249.99.


Weird how so many companies think the first game people will want to play on a new medium is chess. Not only do computers thrash us at it, but fewer people can play it with any competence. Give us mindless sword games with unlimited lives any day.
unique link to this extract

Creating a data set and a challenge for deepfakes • Facebook AI

Mike Schroepfer, chief technology officer:


“Deepfake” techniques, which present realistic AI-generated videos of real people doing and saying fictional things, have significant implications for determining the legitimacy of information presented online. Yet the industry doesn’t have a great data set or benchmark for detecting them. We want to catalyze more research and development in this area and ensure that there are better open source tools to detect deepfakes. That’s why Facebook, the Partnership on AI, Microsoft, and academics from Cornell Tech, MIT, University of Oxford, UC Berkeley, University of Maryland, College Park, and University at Albany-SUNY are coming together to build the Deepfake Detection Challenge (DFDC).

The goal of the challenge is to produce technology that everyone can use to better detect when AI has been used to alter a video in order to mislead the viewer. The Deepfake Detection Challenge will include a data set and leaderboard, as well as grants and awards, to spur the industry to create new ways of detecting and preventing media manipulated via AI from being used to mislead others. The governance of the challenge will be facilitated and overseen by the Partnership on AI’s new Steering Committee on AI and Media Integrity, which is made up of a broad cross-sector coalition of organizations including Facebook, WITNESS, Microsoft, and others in civil society and the technology, media, and academic communities.

It’s important to have data that is freely available for the community to use, with clearly consenting participants, and few restrictions on usage. That’s why Facebook is commissioning a realistic data set that will use paid actors, with the required consent obtained, to contribute to the challenge. No Facebook user data will be used in this data set. We are also funding research collaborations and prizes for the challenge to help encourage more participation. In total, we are dedicating more than $10m to fund this industry-wide effort.


unique link to this extract

YouTubers say kids’ content changes could ruin careers • The Verge

Julia Alexander on the fallout from the FTC nailing YouTube for collecting data about children, and forcing it to stop:


If [YouTube] channels can’t send notifications for certain videos, fewer people will watch those videos within the first crucial hours. This could lead to YouTube recommending fewer videos from that creator because people are less engaged. If videos aren’t recommended as much, it means fewer views, which means less money.

Wojcicki acknowledged that these changes won’t be easy for creators. These changes “will have a significant business impact on family and kids creators,” she said in the post, adding that “this won’t be easy for some creators and are committed to working with them through this transition.”

But creators are coming to terms with exactly how hard it could be. Forrest, a gaming YouTuber with more than 750,000 subscribers who goes by “KreekCraft,” told The Verge that the changes are scary for him. Reading Wojcicki’s blog post only made him feel worse as he tried to figure out, like other YouTube creators, whether his content would be affected by the new system. Would Let’s Play series, tutorials, or even gameplay compilations be considered targeted at children? What’s the difference between family-friendly content and those targeted at kids? No one in the community knows the answers, but everyone is expecting an uphill battle on YouTube under the new system. A YouTube spokesperson pointed The Verge to Wojcicki’s blog when asked for further comment.

“It’s kind of like they’re killing video game content,” Forrest told The Verge. “The top three games on YouTube right now are Fortnite, Minecraft, and Roblox, which are generally non-violent and child-centric games, especially Roblox. Now, we can’t make videos on more mature video games because they’ll get demonetized, but if we make videos on child-friendly games, they’re also now going to get demonetized. What do we do?”


Their problem is that YouTube led them up this path, which turned out to be illegal and unsustainable. The failure is YouTube’s, but it won’t feel it.
unique link to this extract

Graphene-based fabric protects against mosquitoes • Physics World

Sam Jarman:


Graphene-based fabrics could provide an effective new way to protect against mosquitoes according to Robert Hurt and colleagues at Brown University. Using live mosquitoes, the team showed that films of reduced graphene oxide (rGO) are bite-resistant and can block the chemicals that mosquitoes use to detect the presence of skin – even when the material is wet. The group’s insights could provide a basis for new skin coverings that prevent the spread of infectious diseases.

Every year hundreds of millions of people are infected with mosquito-borne diseases such as malaria, dengue and yellow fever – causing about one million deaths worldwide. Preventing mosquito bites therefore plays an important role in public health programmes in many countries.

In recent years, graphene-based materials have been proposed for a wide array of applications, including biomonitoring, sensors, and wearable electronics. Until now, however, protection from mosquito-borne diseases has remained almost entirely unexplored.


Because… it’s really expensive?
unique link to this extract

Beware the Apple iCloud phone phishing scam • Frequent Business Traveler

Anna Breuer:


Scammers have a new and improved way to fool people. A new phone-based phishing scam spoofing Apple’s official support number is likely to take a lot of people by surprise and result in those being called providing the scammers with sensitive information.

The call mimics an official Apple support call, displaying Apple’s logo, Cupertino address, and real toll-free number (800 692-7753). This is the same number, displayed as 800 MY-APPLE, when Apple customers request a call from the company.

Several FBT staffers have reported getting such calls in recent weeks. The calls are not identified by T-Mobile (the mobile operator used by our parent company, Accura) as “Scam Likely” even though it is clear that Apple’s number is being spoofed.

The automated message states that the recipient’s iCloud account “has been compromised” and that he should “stop going online.” The automated message then prompts the caller to dial a toll-free number with an 866 prefix for Apple support.

Typically, Apple’s automated system would prompt the caller to press “1” to be connected to Apple support.

I tried calling the 866 number, which was answered by a main greeting that told me I had reached Apple support and provided an expected wait time. The call was answered by a man with a vague Indian accent who, after asking the reason for my call, disconnected it.


So much excess capacity in Indian call centres; seems like they’ve found a new version of their virus scam.
unique link to this extract

Sonos’ first portable speaker is the $399 Move • The Verge

Dan Seifert:


At almost 10 inches [25cm] tall and weighing over six and a half pounds [3kg], the Move is considerably larger than the Sonos One, making it a bit more to carry around than the typical UE Boom Bluetooth speaker. So Sonos designed a handle directly into the Move’s molded plastic shell to make it easy to pick up and move from room to room or take out of the house. The charging base, which has two pogo pins that line up with the contacts on the back of the Move, give the speaker a “home” when it’s not in use, ensuring it’s charged and ready to go when you need it. If you’re on the go and need to top up the battery, there’s also a USB-C port on the back.

The Move’s larger footprint provides it with more volume and power than the Sonos One. It’s equipped with two Class-D amplifiers, which push a single tweeter and a mid-woofer driver. Sonos says the Move is powerful enough to overcome the rapid falloff in volume that happens when you play music outdoors. The Move also has an IP56 water and dust resistance rating, and the company claims it’s strong enough to withstand accidental falls, rain and moisture, sand and dust, and other elements that might be encountered when a speaker is taken outside of the house.

The Move is also the first Sonos speaker with automatic TruePlay tuning, which lets the speaker adapt its sound for its environment. With earlier Sonos speakers, TruePlay tuning required walking around a room with an iPhone or iPad while a beeping tone played from the speaker to “map” the room. The Move can use its own microphones to adjust its sound within about 30 seconds of playback, which is much easier than the prior method and convenient for a speaker that will migrate from place to place on a regular basis.


One beta tester told me “it weighs a ton!” That’ll be the battery. Life is quite a challenge for Sonos, which is facing disruption below from cheap Bluetooth speakers, and competition alongside from Amazon and Google, and kinda from above from Apple. Its best hope is being the cross-platform solution that plays nicely with all of them. But: not cheap. $399 in the US, £399 in the UK.
unique link to this extract

Samsung and EE bring Galaxy Fold 5G to the UK • Samsung Newsroom U.K.


Samsung Electronics Co., Ltd. has today announced that the Galaxy Fold 5G will be available to buy from 18th September in the UK via an exclusive operator partnership with EE, as well as from Samsung Experience Stores. The device will also be displayed at Samsung KX, Harrods and Selfridges for customers to experience.
The Galaxy Fold 5G, which will be available in Cosmos Black and Space Silver, pushes the boundaries of innovation and introduces a whole new smartphone category. Armed with 5G network capabilities, the Galaxy Fold 5G is a device built for the future…

…The Samsung Galaxy Fold 5G will be available from Samsung at an RRP of £1,900 and all devices will come with wireless Galaxy Buds and a Galaxy Fold 5G Aramid case. EE price plans will be announced in due course.


EE doesn’t offer any Sim-only 5G plans, so it’s impossible to say what extra you might be paying annually. EE offers seven 5G phones, with the cheapest being £44 per month for a refurbished Galaxy S10.

For comparison, the Galaxy Note10+ 5G costs £1,099 for the 256GB model (with no network connectivity). EE wants £84 per month for unlimited text, data and talktime at 5G – but it doesn’t say how long the contract lasts. 12, 18, 24 months? It’s never specified. Let me know if you find out. A 12-month contract would cost £1,008; an 18-month one, £1,512. A 24-month one (which I suspect it is) would be £2,016. Also, the price would rise by inflation (RPI) every March. As ever, it’s better to buy the phone and get a Sim.
unique link to this extract

Why ‘SIM swapping’ is a growing security nightmare • The New York Times

Nathaniel Popper:


“I’ve been looking at the criminal underground for a long time, and SIM swapping bothers me more than anything I’ve seen,” said Allison Nixon, the director of research at the security firm Flashpoint. “It requires no skill and there is literally nothing the average person can do to stop it.”

Criminals have learned how to convince mobile phone providers like T-Mobile and AT&T to switch a phone number to a new device that is under their control.

The number is switched from a tiny plastic SIM card, or subscriber identity module, in the target’s phone to a SIM card in another device.

Sometimes hackers get phone numbers by calling a customer help line for a phone carrier and pretending to be the intended victim. In other recent incidents, hacking crews have paid off phone company employees to do the switches for them, often for as little as $100 for each phone number.

Once the hackers have control of the phone number, they ask companies like Twitter and Google to send a temporary login code, via text message, to the victim’s phone. Most major online services are willing to send those messages to help users who have lost their passwords.

But the temporary code is sent to the hackers.

Phone companies have been aware of the problem for years, but the only routine solution they have come up with is offering pin codes that a phone owner must provide in order to switch devices. Even this measure has proved ineffective. Hackers can get the pin codes by bribing phone company employees.


Personally, I don’t use two-factor systems that send phone codes, if at all possible. Even Twitter has finally – finally! – moved to a system where the 2FA can rely on a time-limited code generated by an app.
unique link to this extract

Drone bubble bursts, wiping out startups and hammering VC firms • Bloomberg

Jack Pitcher:


Once well-funded startups are struggling as hordes of self-employed pilots drive down prices, Chinese technology races ahead and non-drone companies across industry pull their unmanned aerial operations in-house. Federal regulation of the aircraft has been slow to catch up, and is holding back many businesses from expanding.

French manufacturer Parrot SA announced in July that it would halt production of most of its drone lines. Software startup Airware raised $118m from investors before shutting its doors and laying off 140 employees in late 2018. GoPro exited the drone business and laid off hundreds last year, citing an “extremely competitive” market.

But while some startups are testing investor patience, others are seeing an opportunity for growth. At least 67 drone startups have been sold since their inception, according to Crunchbase, which collects data on private companies. Buyers range from rival drone operators to companies in other industries, such as Verizon Communications…

…Venture capitalists poured $2.6bn into drones from the beginning of 2012 to June 2019, according to Teal Group, an industry researcher. The rapture began to evaporate last year as startups founded during ‘peak hype’ in the commercial drone industry ran out of money before they could generate profit and couldn’t secure additional funding, said Wackwitz.

At least 25 drone startups have shut their doors this decade, with the largest burning through a total of $183m in funding, according to Crunchbase’s online reports.

“The venture capitalists are less enthused now,” said Dan Burton, CEO of Dronebase, a drone pilot network that’s held on through the turmoil.


Gee, ya think? But it does illustrate how what seems like an absolute slam-dunk of a market – hey, we can take pictures from way up high! – turns out to have a seriously limited addressable market. Films and TV use drones regularly, farmers do, planners might, but those billions invested were probably 10x the total market size.
unique link to this extract

Errata, corrigenda and ai no corrida: none notified

6 thoughts on “Start Up No.1,139: how bots can change votes, the drone bust, pricing the Galaxy Fold, Sonos gets mobile, and more

  1. 2FA that relies on a time-limited code generated by an app is great until your phone breaks. Then it can turn into a bit of a problem as I’ve recently discovered (it’s going to take me ages to get everything back as the certificates for all the authenticators doesn’t appear to be in the backup when you restore).

      • An app is only as safe as the subcontractor’s subcontractor’s trainee is incorruptible. I know I should use one, but even before the stories ( ), Password Managers feel like very high-stakes targets for government and mafia hackers.

        Even if the code is perfectly secure at one point, even if the passwords are encoded with your own key, it only takes one hacked version of the app pushed to your device for all that security to become an illusion. The basket is supposedly secure, but putting all my eggs in a basket that’s crafted by hundreds of people and swapped on me at will…

        Granted, same is true for Android, iOS, Google, and Apple. But at least I’m not adding one extra layer of vulnerability in the form of a much smaller corp ? I’m not even sure a smaller corp is more vulnerable, but if I can do w/o it…

    • When you set up a service’s (app, site, server…) 2FA in your 2FA app, the source displays a barcode for the 2FA app to scan. You can, et that moment and that moment alone, scan this code on several devices (someone pointed out you can also screenshot it, but that’s less safe; you can also printout one-time codes, again, less safe).

      I did that and have 2FA apps filled with codes on my main phone, my backup phone, and my main tablet. Behind MIUI’s “secure app” protection, so those 2FA apps request my fingerprint to start.

      The cheapest Xiaomi is $90 (pin only, $30 extra for fingerprint ID, frequent 20%ish promos), runs OK, even has that fancy Dual Spaces stuff for when you loan it out. Between the safety on a second 2FA box and the practicality of a backup/second phone (doesn’t even need to have a SIM/contract), it’ll pay for itself. Add another $30 and get a Redmi Note ;-p

  2. I’d say drones need their Apple moment. Right now they’re neither easy nor sexy, so only nerds can use them, and since they’re nerds, their main use for a drone is… to use a drone. Also, pro photographers – but the potential for abuse is so high that the permits required are costly and kafkaesque (in my windy Provence, permits are not deferrable, so if you got wind and can’t shoot, you got to re-apply for a new permit, which takes 2-3 weeks, ie outside of weather forecasts).

    If drones were a) silent b) intelligent (about both coming back for juice and framing pics), things would be different. I could send it to “take a pic of Etienne as he practices his ninja dives” or “take a group shot”; maybe it could have some Google Clip smarts and just hover about shooting everything and keeping the good stuff. Or coordinate a flotilla for events (mkjing sure to shoot both everyone and everything).

    We’ll probably get that eventually. But drones might have been Glass’d already; I certainly think of most users as dronholes.

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 )

Twitter picture

You are commenting using your Twitter 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.