The FBI finds these unsettling. CC-licensed photo by Steve Garfield 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 9 links for you. Non-algorithmic. I’m @charlesarthur on Twitter. Observations and links welcome.
“F**k the algorithm?”: what the world can learn from the UK’s A-level grading fiasco • Impact of Social Sciences
Last month, hundreds of students in UK gathered in front of the Department for Education and chanted “f**k the algorithm”. Within days, their protests prompted officials to reverse course and throw out test scores that an algorithm had generated for students who never sat their exams due to the pandemic.
This incident has shone the media spotlight on the question of AI bias. However, previous cases of AI bias have already led to well-intentioned efforts by data scientists, statisticians, and machine learning experts to look beyond the technical and also consider the fairness, accountability, confidentiality, and transparency of their algorithms.
What the A-level grading fiasco demonstrates is that this work may be misdirected. There is a key lesson to be learned from this algorithmic grading fiasco. A lesson that will only become more relevant as governments and organizations increasingly use automated systems to inform or make decisions: There can be no algorithmic accountability without a critical audience. By this, I mean that, unless it draws the attention of people who critically engage with it, technical and non-technical quality assurance of algorithms is a token gesture and will fail to have the desired effect.
During the Epstein mania, for example, #ClintonBodyCount trended, and President Trump — who is known to trawl Twitter trends for material — retweeted an account that sought to link Epstein’s death to the former president.
To me, the incident offered reasons to bring an end to trending topics altogether. One, the Twitter was seemingly powerless to stop bad actors from gaming the algorithm and inserting fringe ideas into mainstream discussion. And two, the feature had been made largely redundant by Moments, a 5-year-old product that uses human curators to find items of interest on Twitter each day and organize meaningful discussions around them.
In conversations with Twitter around that time, executives told me that they knew their trends had problems, but ensured me that fixes were coming. On Tuesday — more than a year after the disinformation World Cup — the first such fix arrived. The company announced it in a blog post:
Sometimes the right Tweet can help make sense of a trend. Starting today, some trends will have a representative Tweet pinned to them to give you more insight about a trend right away.
…now when something trends, you’ll see a tweet that explains why, plus maybe a short explanation from Twitter. If nothing else, this should resolve what might be the most common complaint about trends for the past decade or so: whenever a celebrity’s name is trending, everyone assumes they are dead, and has to frantically search through tweets to see whether that is in fact the case.
…Reading through the list of conspiracy topics that have surfaced in Twitter trends over the past year, it’s hard to imagine how the changes announced on Tuesday will much improve the product. Will trends be worthy when a human curator picks a “representative tweet” for, uh, #JewishPrivilege? What about #SaveTheChildren?
Most of us know how delightful it is to hear a computer-generated song playlist that feels entirely personal. Now, Google wants to create a similar type of bespoke audio experience—not with music, but with news.
The company is adding some new features to its existing news aggregation service called Your News Update, which gathers news clips from different outlets and plays them in one continuous audio feed. Think of it like a Feedly or Flipboard-type service for spoken stories from your preferred news publications.
Google has updated the service to create a more fluid listening experience, so that sitting through an entire session doesn’t feel like you’re just working your way through a hodgepodge of disparate stories. Each personalized playlist is structured to mimic a news program typical of what you’d hear on public radio: short clips about the big headlines up front that gradually shift into longer, more detailed stories. The goal is to create a seamless 90-minute broadcast—a mix of radio, podcast snippets, and text-to-speech article translations—tailored to an audience of one.
Or you could listen to the news, which would give you information about the whole world and take you out of your filter bubble? This is one of those ideas that must sound brilliant in the product meeting, but which doesn’t actually have good outcomes if adopted widely.
unique link to this extract
once Covid-19 has established itself in the body, things start to get really interesting. According to Jacobson’s group, the data Summit analyzed shows that Covid-19 isn’t content to simply infect cells that already express lots of ACE2 receptors.
Instead, it actively hijacks the body’s own systems, tricking it into upregulating ACE2 receptors in places where they’re usually expressed at low or medium levels, including the lungs.
In this sense, Covid-19 is like a burglar who slips in your unlocked second-floor window and starts to ransack your house. Once inside, though, they don’t just take your stuff — they also throw open all your doors and windows so their accomplices can rush in and help pillage more efficiently.
The renin–angiotensin system (RAS) controls many aspects of the circulatory system, including the body’s levels of a chemical called bradykinin, which normally helps to regulate blood pressure. According to the team’s analysis, when the virus tweaks the RAS, it causes the body’s mechanisms for regulating bradykinin to go haywire. Bradykinin receptors are resensitized, and the body also stops effectively breaking down bradykinin. (ACE normally degrades bradykinin, but when the virus downregulates it, it can’t do this as effectively.)
The end result, the researchers say, is to release a bradykinin storm — a massive, runaway buildup of bradykinin in the body. According to the bradykinin hypothesis, it’s this storm that is ultimately responsible for many of Covid-19’s deadly effects.
The Indian government has banned 118 more Chinese mobile apps in the country, including mega-popular action title PUBG Mobile.
In a press release explaining the move, India’s Ministry of Electronics and Information Technology said they chose to ban the apps as they “are engaged in activities which is prejudicial (sic) to sovereignty and integrity of India, defence of India, security of state and public order.”
More specifically, the ministry claims that it received many complaints about these mobile apps “stealing and surreptitiously” sending user data to servers outside of India.
Aside from PUBG Mobile, some of the other notable names in this latest list of banned apps includes PUBG Mobile Lite, AFK Arena, AliPay, Baidu, several WeChat-branded apps, APUS Launcher, and Rules of Survival.
Not much noticed on this side of the world, but the cold war going on between India and China is pretty intense.
unique link to this extract
Alex Hern on the impasse in Australia, under which Facebook says it will stop including news links because of a new payment requirement:
One of the problems with technology regulation is that “that would be bad”, “that would be bad for tech” and “that would be bad for tech companies” are often treated as interchangeable pieces of advice.
One solution is to build up your technical expertise so that you can distinguish between those different types of advice, learning from the experts where they are helpful while dismissing their input when it is self-serving.
Another solution – the Australian federal solution – is to ignore the experts entirely and throw yourself over the waterfall in a barrel. Maybe you’ll be fine! Maybe you won’t be! Either way, everyone watching will have a grand old time, and we’ll probably learn something in the process.
… one thing I have liked about the Australian approach, as I so frequently do, is the fact that completely ignoring the experts does at least highlight some of the problems with the expert advice.
…Facebook’s point is that if the Australian news companies want to charge it per (say) headline on Facebook, while also retaining the ability to post their own headlines on Facebook… the company is signing a blank cheque. I can see why it would be against that!
Grab your phone, open Instagram, click add to Your Story, and flick over to Reels. You’ll see a little musical note. Click that, and you can search for almost any song you like, and add it to your Reel.
It’s a feature lifted wholesale from TikTok, of course, but that’s not important here; what is important is that it’s a feature for which Facebook has, almost certainly, negotiated a small fee per stream to be paid back to the labels.
In other words, Facebook has signed a blank cheque in order to allow excerpts of music be posted on Instagram
Throw yourself over the waterfall in a barrel. That’s 2020 in a nutshell. Or a barrel.
unique link to this extract
Google has told its tens of thousands of clients that from November it will charge an additional fee for ads served on Google and YouTube. The move will increase advertisers’ costs in line with the amount the tech giant is set to pay in new digital services taxes as they come into force: 2% in the UK, and 5% in Austria and Turkey.
Earlier this month, Amazon said that it would be passing on the 2% tax to sellers on its platform. With Google doing the same, Facebook, which makes an estimated £4.2bn in ad revenue in the UK, is expected to follow suit.
“Digital service taxes increase the cost of digital advertising,” said a Google spokeswoman. “Typically, these kinds of cost increases are borne by customers and like other companies affected by this tax, we will be adding a fee to our invoices, from November. We will continue to pay all the taxes due in the UK, and to encourage governments globally to focus on international tax reform rather than implementing new, unilateral levies.”
Google UK reported £1.6bn in revenues last year, up from £1.2bn, but paid just £44m in UK corporation tax as it does not report big profits.
However, this does not reflect how much it makes in total advertising revenues as they are reported in other jurisdictions. The research firm eMarketer estimates that in reality Google made about £5.7bn in ad revenue in the the UK last year, accounting for 39% of that total digital ad market, and will make over £6bn this year. The 2% tax equates to about £122m on those revenues.
Apple following suit by increasing the cost of its developer accounts by 2% in the UK (before it puts on the 20% VAT). Not sure that’s how taxes are meant to play out.
unique link to this extract
The downside for police, who have rushed to embrace Ring usage nationwide as the Amazon subsidiary aggressively marketed itself to and sealed partnerships with local departments, is that networked cameras record cops just as easily as the rest of us. Ring’s cameras are so popular in part because of how the company markets their ability to detect motion at your doorstep, providing convenient phone alerts of “suspicious activity,” however you might define it, even when you’re out of the house.
But sometimes the police are the unannounced, unwanted visitor: “Subjects likely use IoT devices to hinder LE [law enforcement] investigations and possibly monitor LE activity,” the bulletin states. “If used during the execution of a search, potential subjects could learn of LE’s presence nearby, and LE personnel could have their images captured, thereby presenting a risk to their present and future safety.”
The document describes a 2017 incident in which FBI agents approached a New Orleans home to serve a search warrant and were caught on video. “Through the Wi-Fi doorbell system, the subject of the warrant remotely viewed the activity at his residence from another location and contacted his neighbor and landlord regarding the FBI’s presence there,” it states.
Sauce for the goose turns out to serve gander as well.
unique link to this extract
Intel has officially announced its first 11th Gen Tiger Lake processors for laptops, which will feature the company’s new integrated Xe graphics, Thunderbolt 4 support, Wi-Fi 6, and a big leap in performance and battery life over the previous Ice Lake chips. The company claims that the new 11th Gen lineup offers the “best processor for thin-&-light” laptops.
Intel is launching nine new 11th Gen designs for both its U-series (which Intel is now referring to as UP3) and Y-series class chips (aka UP4), led by the Core i7-1185G7, which offer base speeds of 3.0GHz, a maximum single core turbo boost of up to 4.8GHz, and a maximum all-core boost of up to 4.3GHz. It also features the most powerful version of Intel’s Iris Xe integrated graphics, with 96 CUs and a maximum graphics speed of 1.35GHz.
…Intel isn’t being too specific on what those increases will be, but it promises that the new chips will offer a 20% faster speeds for day-to-day “office productivity” tasks, along with a similar 20% increase in “system-level power,” which is says results in more than an extra hour of battery life for things like video streaming.
…Also new is support for 8K HDR displays, along with the option to use up to four 4K HDR displays at once. There’s also improvements to the built-in AI engine, which Intel says will offer specific improvements for video calls (like background blurring) — tasks which ARM-based computers like the Surface Pro X have previously excelled at.
Gauntlet thrown down to Apple. It’s a bold strategy, Cotton, let’s see if it pays off for ’em.
unique link to this extract
Errata, corrigenda and ai no corrida: none notified