Start Up No.1830: explaining those odd wrong-number texts, Covid’s missing immunity, Google offers abortion data deletion, and more

Though humans struggle to understand what baby chickens are expressing, machine learning systems can figure it out. CC-licensed photo by Nenad Stojkovic 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. Use them wisely. I’m @charlesarthur on Twitter. Observations and links welcome.

What’s the deal with all those weird wrong-number texts? • Substack

Max Read:


This [strange “misdirected” text/WhatsApp/message] is the first step in what is, at its core, an old-fashioned “romance scam,” in which the scammer exploits a lonely and/or horny person by faking a long-distance, usually romantic relationship. After the scammer has gained the trust of their victim, they convince them to transfer money, often for an investment; in some cases, the victim can be enticed into several successive transfers before they realize they’re being played.

This kind of con has proliferated over the last few years in China, where it’s called sha zhu pan, or “pig-butchering,” because the victim is strung along for weeks or months before the actual swindle, like a pig being fattened for slaughter. Originating in sophisticated online-fraud networks first developed to take advantage of Chinese offshore gamblers, the sha zhu pan scams end with targets depositing money into forex or gold trading — or, seemingly most commonly, into fake cryptocurrency platforms. (Interestingly, they’re often not “romantic” at all, and instead rely on cultivating a trusting friendship that culminates with a little bit of friendly investing advice.)

While sha zhu pan scams are common enough in and around China that there are Chinese-language YouTubers whose stock in trade is identifying and publicizing scams, the same scam networks seem to have expanded outward over the past year or so, joining America’s (and Europe’s) own homegrown romance and crypto-scam industries on dating sites and, yes, via “accidental” wrong number texts.


The selection of “oops misdirected” scam opener texts is amazing. I think “Andy, will my custom mahogany furniture arrive next week?” is my favourite. It’s a terrific read.
unique link to this extract

Where’s the herd immunity? Our research shows why Covid is still wreaking havoc • The Guardian

Danny Altmann is a professor of immunology at Imperial College London:


During May and June two new variants, BA.4 and BA.5, progressively displaced the previous Omicron subvariant, BA.2. They are even more transmissible and more immune-evasive. Last week a group of collaborators, including me and a professor of immunology and respiratory medicine, Rosemary Boyton, published a paper in Science, looking comprehensively at immunity to the Omicron family, both in triple-vaccinated people and also in those who then suffered breakthrough infections during the Omicron wave. This lets us examine whether Omicron was, as some hoped, a benign natural booster of our Covid immunity. It turns out that isn’t the case.

We considered many facets of immunity, including the antibodies most implicated in protection (“neutralising antibodies”), as well as protective “immune memory” in white blood cells. The results tell us it is unsurprising that breakthrough infections were so common. Most people – even when triple-vaccinated – had 20 times less neutralising antibody response against Omicron than against the initial “Wuhan” strain. Importantly, Omicron infection was a poor booster of immunity to further Omicron infections. It is a kind of stealth virus that gets in under the radar without doing too much to alert immune defences. Even having had Omicron, we’re not well protected from further infections.

Also, to be added to the now complex mix is “immune imprinting”. This is the finding that our immune response to Covid is shaped very differently, depending on our prior exposures – infection in one wave relative to another, plus vaccination. In our study, those who’d been infected in the first wave and then again with Omicron had particularly poor T-cell responses and no boosting of antibodies. That is, some combinations of exposures may leave us poorly protected relative to others.

Contrary to the myth that we are sliding into a comfortable evolutionary relationship with a common-cold-like, friendly virus, this is more like being trapped on a rollercoaster in a horror film. There’s nothing cold-like or friendly about a large part of the workforce needing significant absences from work, feeling awful and sometimes getting reinfected over and over again, just weeks apart. And that’s before the risk of long Covid.


unique link to this extract

Artificial intelligence could spot baby chickens in distress • Science

Virginia Morell:


Early in life, chicks utter distress calls—high-pitched, repetitive chirps—to attract the attention of their mother hen, whom they rely on for warmth and food. She responds with food calls, showing the chicks where to forage. But in a commercial chicken barn, chicks call out when they’re uncomfortable, socially isolated, or hungry. Answering these calls can be the difference between life and death: Ignored chickens are more likely to lose weight and die prematurely. Animal welfare scientists have been trying to develop automatic methods to help farmers better spot these situations.

To improve these efforts, researchers at the City University of Hong Kong recorded the vocalizations of chickens housed at Lingfeng Poultry Ltd., a major poultry producer in China’s Guangxi province. The birds are kept in stacked cages (three cages per stack, and 13 to 20 individuals per cage), with about 2000 to 2500 chickens in each barn.

Over the course of a year, the researchers recorded the environment, picking up everything from natural farm sounds such as workers hosing down barn floors to the chick distress calls. They then transformed all of these noises into sound pictures known as spectrograms and used the images to train a type of AI program called deep learning. Similar programs have been developed to recognize the emotional states of cows on dairy farms.

Using the recorded sounds from the barns as well as sounds made in real time in a live demonstration, the algorithm rapidly and successfully identified 97% of distress calls as the chickens were making them, distinguishing these from other chicken sounds and from general barn noise, the team reports today in the Journal of the Royal Society Interface.


Cor blimey guvnor it’s a right old Doctor Dolittle. What a strange possibility that machines might be able to understand animal communications better than us and infer their thoughts and desires. (Think we’ll be ahead with dogs for a while though.)
unique link to this extract

Google will delete location history data for abortion clinic visits • Reuters via The Guardian


Google will delete location data showing when users visit an abortion clinic, the online search company said on Friday, after concern that a digital trail could inform law enforcement if an individual terminates a pregnancy illegally.

As state laws limiting abortions set in after the US supreme court decided last month that they are no longer guaranteed by the constitution, the technology industry has fretted police could obtain warrants for customers’ search history, geolocation and other information revealing pregnancy plans.

Google on Friday said it would continue to push back against improper or overly broad demands for data by the government, without reference to abortion.

The company said the location history of a Google account was off by default.

Effective in the coming weeks, for those who do use location history, entries showing sensitive places including fertility centers, abortion clinics and addiction treatment facilities will be deleted soon after a visit.

A Google spokesperson did not immediately answer how the company would identify such visits or whether all related data would be wiped from its servers.


That last part is the tricky one. How will it identify the visits? And the bit about the location history being “off by default” doesn’t really fit with the way that Android hassles you for access to your location.
unique link to this extract

New York denies air permit to bitcoin mining power plant • The Verge

Justine Calma:


Bitcoin miners in New York state faced a regulatory blow last week as the state denied air permits for a gas-fired power plant used to mine bitcoin. It’s the latest step that New York has taken to crack down on crypto mining as it tries to meet its goals on climate change.

The decision was made for the Greenidge Generating Station in New York’s Finger Lakes region. Bitcoin mining brought new life and renewed controversy to the embattled plant in 2020. That drew outrage from some local residents worried about how the plant could affect fish and tourism by discharging hot water into nearby Seneca Lake. At the state level, Greenidge’s revival has sparked fears that pollution from the energy-intensive process of mining Bitcoin could revive other zombie power plants and derail New York’s climate goals.

New York state set a goal in 2019 of slashing its greenhouse gas emissions by at least 85% over the next few decades. The fight over Greenidge has been billed up as a test of how serious the state is about meeting that goal. Is it willing to get tough on the lucrative bitcoin industry that’s boomed in New York ever since China kicked out miners last year?

…Greenidge operated as a coal-fired power plant for decades. But as coal struggled to compete with cheap natural gas across the country, the plant temporarily shuttered before retrofitting itself to run on gas in 2017. Then, in 2020, the plant’s operators spotted a more lucrative venture and started mining Bitcoin, which now makes up the vast majority [over 95%] of the company’s revenues.


With the price cratering, hard to see the permits and the hassle being worth it.
unique link to this extract

Former top Apple lawyer pleads guilty to insider trading • Reuters via CNBC


The former top corporate lawyer at Apple pleaded guilty on Thursday to insider trading charges, for what prosecutors called a five-year scheme to trade ahead of the iPhone maker’s quarterly earnings announcements.

Gene Levoff, 48, of San Carlos, California, pleaded guilty to six securities fraud charges at a hearing before US District Judge William Martini in Newark, New Jersey. [Apple fired Levoff in September 2018, five months before he was criminally charged.]

Levoff allegedly exploited his roles as corporate secretary, head of corporate law and co-chair of a committee that reviewed drafts of Apple’s results to generate $604,000 of illegal gains on more than $14m of trades from 2011 to 2016.

Prosecutors said Levoff ignored the quarterly “blackout periods” that barred trading before Apple’s results were released, as well as the company’s broader insider trading policy — which he was responsible for enforcing.

“Gene Levoff betrayed the trust of one of the world’s largest tech companies for his own financial gain,” First Assistant US Attorney Vikas Khanna in New Jersey said in a statement.


Could get up to 20 years and $5m fine per count, though probably won’t. You can see why he wasn’t worried about the person watching out for insider trading spotting him.
unique link to this extract


Fun game: press keys on the keyboard to get a different sound from each of a-z. Press the spacebar to get a different set of sounds. Not clear that you can record them, but a fun way to pass the time. Also available as an iOS app.
unique link to this extract

Brown Bear Cam – Brooks Falls in Katmai National Park •

You wanted webcams showing animals in natural and not-so-natural settings? Here’s a page full of them. (Includes sleeping dogs, cats, fish, etc.) Beats the days when all you could see was a pot of coffee.
unique link to this extract

Meta’s shutting down its digital wallet, Novi • The Verge

Emma Roth:


Meta’s ending the pilot for Novi, the company’s digital wallet and the last remaining piece of its troubled cryptocurrency project, as first reported by Bloomberg. On Novi’s website, Meta says the wallet is shutting down on September 1st, 2022, and asks users to withdraw their funds “as soon as possible.”

Users will lose access to their accounts come September, and will no longer be able to add money to Novi starting July 21st. If someone forgets to withdraw their remaining balance, Meta says it will “attempt to transfer” their funds to the bank account or debit card added to the service.

Meta rolled out the “small pilot” of Novi to users in the US and Guatemala last October. Novi was originally built to support fast and free transactions using the Meta-backed cryptocurrency, Diem, but regulatory challenges forced the company to partner with Coinbase to use the Paxos stablecoin (USDP) instead. While Meta made it clear that it still planned on adding support for Diem at a later date, things started to fall apart (more than they already were) at the end of 2021 and into 2022.

Before Facebook’s parent company was known as Meta, Diem was also known by another name: Libra. The cryptocurrency project faced scrutiny over its ties to Facebook, so much so that the independent group behind Libra rebranded the project to Diem in an attempt to distance itself from the social network.

Members of the US Senate called on Meta to shut down its Novi project shortly after its October 2021 launch, citing that the company “cannot be trusted to manage cryptocurrency.”


My concern about Facebook and Libra (as it was) was the idea that it could become a de facto global currency, overseen by Mark Zuckerberg. You only have to reflect on that idea – of transactions happening within Facebook because they’re more convenient, and all the arbitrage between currencies being handled at the back end, until Facebook is the one dictating exchange rates – to think the outcomes aren’t good.
unique link to this extract

• Why do social networks drive us a little mad?
• Why does angry content seem to dominate what we see?
• How much of a role do algorithms play in affecting what we see and do online?
• What can we do about it?
• Did Facebook have any inkling of what was coming in Myanmar in 2016?

Read Social Warming, my latest book, and find answers – and more.

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 )

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.