Start Up No.1498: Google says it will stop ad tracking, the deepfake TikTok Tom Cruise, does vaccination cure long Covid?, and more

Anyway, the CIA wanted to put a plutonium-powered sensor atop this mountain. Instead, it’s probably in a glacier. CC-licensed photo by Anirban Biswas 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. Use them wisely. I’m @charlesarthur on Twitter. Observations and links welcome.

Google to stop selling ads based on your specific web browsing • WSJ

Sam Schechner and Keach Hagey:


Google plans to stop selling ads based on individuals’ browsing across multiple websites, a change that could hasten upheaval in the digital advertising industry.

The Alphabet company said Wednesday that it plans next year to stop using or investing in tracking technologies that uniquely identify web users as they move from site to site across the internet.

The decision, coming from the world’s biggest digital-advertising company, could help push the industry away from the use of such individualized tracking, which has come under increasing criticism from privacy advocates and faces scrutiny from regulators.

Google’s heft means that its move is also likely to stoke a backlash from some competitors in the digital ad business, where many companies rely on tracking individuals to target their ads, measure their effectiveness and stop fraud. Google accounted for 52% of last year’s global digital ad spending of $292bn, according to Jounce Media, a digital-ad consultancy.

“If digital advertising doesn’t evolve to address the growing concerns people have about their privacy and how their personal identity is being used, we risk the future of the free and open web,” David Temkin, the Google product manager leading the change, said in a blog post Wednesday.


The blogpost is unequivocal:


Today, we’re making explicit that once third-party cookies are phased out, we will not build alternate identifiers to track individuals as they browse across the web, nor will we use them in our products.


This is jawdropping stuff, and it’s impossible not to wonder: what’s Google’s real plan here? It’s like the tale of the French diplomat who was famous for his subtle, clever dealings; when he died, one rival said “hmm, I wonder what he intended by that?”

Obvious questions: does this mean that all the insistence on tracking (“to serve you relevant ads”) was wasted, and isn’t worth doing? If not, how will Google keep on doing tracking? Or if it wasn’t wasted and they aren’t going to keep on doing the tracking, how will they make the ads pay as well as they used to?

The timing – just ahead of Apple introducing its IDFA ad-tracking-opt-in requirement in iOS 14.5 – is surely not accidental. Google may have been working on this for some time, and realised that 14.5 will put a lot of rivals at a disadvantage, and is getting out ahead of it.
unique link to this extract



Consumers deserve a data dividend • Oracle corporate blog

Ken Glueck is executive VP at Oracle:


consumers are shamefully under-compensated for the data they generate as they go about their daily life, both virtual and physical. Consumers are told that the trade of their personal data in exchange for “free” services is a fair one. It is not. The marginal cost of any of these services is effectively zero yet the ad revenue generated from consumer data easily exceeds $200bn. On the other side of the market Google’s privacy policy and terms of service guarantees its cost to acquire consumer data is zero. It’s a taking. The only one actually getting something for free is Google.

And that’s not all. The most subtle unfairness of this exchange is that consumers never stop paying, and the amount of data taken from them keeps rising. The data extracted from consumers now wildly exceeds the value of what they receive. What’s worse, Google’s data collection is increasingly untethered from a user’s underlying activity. For instance, on an Android phone, Google collects a range of location and sensor information irrespective of what services or apps the user has active, without their direct action, via processes hidden in the background.

Much of this consumer data is not, or should not, be Google’s to take in the first place. The “consent” for this data collection comes from the “notice” provided by “privacy policies,” which in turn embeds Google’s all-encompassing terms of service. So, not unlike the shrink-wrap licenses of the 1980’s, Google’s click-wrap licenses—or contracts of adhesion—give Google cart blanche to collect whatever user data it desires, whenever it desires. Boiled down to its simplest terms, the consideration consumers receive for their highly valuable stream of data is worth a lot more than what Google is providing.


Why, you might wonder, is Oracle (jilted suitor of TikTok) griping about Google? Because it has a long-running lawsuit, presently being considered by the US Supreme Court, accusing Google of breaking the copyright around the APIs for Java. And while it waits for that judgment, it might as well snipe at Google in other ways.

Wildly ironic, considering the link before, that Oracle chose today, of all days, to post this.
unique link to this extract



Tom Cruise isn’t on TikTok: it’s a shockingly-realistic deepfake • PetaPixel

Jaron Schneider:


A TikToker is using deepfake technology to impersonate Tom Cruise on the social media platform and the results are so realistic that some may mistakingly believe it actually is the famed actor. This latest situation has again raised concerns about the creation and use of deepfakes.

While the account is clearly making folks aware that this isn’t the real Tom Cruise — the username is deeptomcruise, for starters– those not paying attention can easily mistake what they’re seeing for the genuine article. Even without seeing the username, the video isn’t quite perfect (The Verge notes that the lip-syncing is off in places and the voice isn’t quite right).

The most recent video, uploaded four days prior to publication, is the most realistic of the batch and depicts the Cruise impersonater performing a magic trick.

Again, looking closely reveals that something is amiss, but no doubt this video would fool many and it’s clearly close enough to raise the alarm as multiple publications have weighed in on the account that is once again causing some to question the legality of deepfakes.

Overall, the account has more than 10 million views, 1.1 million likes, and over 370,000 followers. On Tuesday afternoon, coverage of the account reached a fever pitch and was trending on Twitter.

According to TikTok’s own terms of service, the Tom Cruise impersonation videos should be a violation:


You may not: […]
impersonate any person or entity, or falsely state or otherwise misrepresent you or your affiliation with any person or entity, including giving the impression that any content you upload, post, transmit, distribute or otherwise make available emanates from the Services


Yet days after the initial story of the account’s viral spread broke, the videos remain on the platform.


Seem to have been taken down now. They’re really, really impressive. We’ve come a long way in the three years since the first deepfake videos began appearing on porn sites.
unique link to this extract



The dark legacy of the CIA’s bungled plot to have famous climbers plant nuclear-powered sensors in the Himalayas • Defector

Patrick Redford:


China’s nuclear test announced the country as the world’s fifth nuclear-armed nation, though unlike France, the USSR, and the United Kingdom, the Chinese nuclear program was a black box for American intelligence. Recently declassified government files show, for example, that the U.S. was shocked to learn that the bomb was fueled by uranium, not plutonium. Military-industrial honchos were left scratching their heads as to how to gather intelligence, until a chance meeting between General Curtis LeMay and mountaineer Barry Bishop at a Washington D.C. cocktail party led to one of the most quixotic, unsuccessful operations in the CIA’s long history of screwups.

Bishop was part of the first American team to summit Mt. Everest the year prior, and according to Pete Takeda’s fascinating 2007 Rock And Ice story, he gushed about the unobstructed views he enjoyed from the world’s (arguably) tallest mountain. Takeda writes that LeMay put the pieces together and, “From this casual exchange emerged an unlikely inspiration: Recruit America’s best high-altitude climbers to place a nuclear powered observation device atop the world’s greatest mountain range.” The hope was that a transceiver could pick up radio communications between Chinese nuclear personnel, remaining functional for years off of the heat from decaying plutonium isotopes. Per Takeda, the CIA’s device was an “oven-sized metal bin with five radiating fins” that weighed 125 pounds and was topped by a six-foot long antenna. If this sounds like a crude product of ’60s nuclear frenzy, consider that NASA’s Perseverance rover is scooting around on Mars thanks to this exact sort of battery.


There’s also a writeup over at Rock & Ice (which this piece borrows heavily from, but has the better headline) going into some more detail about what might have happened to the device. A clue: nothing good, especially if you live near a glacier at the bottom of Nanda Devi.
unique link to this extract



Long COVID patients say they feel better after getting vaccinated • The Verge

Nicole Wetsman:


“I started getting texts and calls from some of my colleagues saying hey, are your patients with long COVID reporting that they’re feeling better after the vaccine?” says [Daniel] Griffin, an infectious diseases clinician and researcher at Columbia University. When he started talking with patients, he saw that they were. “It’s not 100 percent, but it does seem like to be around a third,” he says.

Early reports from Griffin and others hint that people with persistent symptoms may improve after getting vaccinated. Information is still limited, and the data is largely anecdotal — but if the pattern holds, it could help researchers understand more about why symptoms of COVID-19 persist in some people, and offer a path to relief.

Many of Griffin’s patients who improved had significant side effects after their first shot of either the Moderna or Pfizer / BioNTech vaccine. That’s common in people who’ve had COVID-19 before — they already have some level of antibodies, so the first shot acts more like a second booster. Then, his patients with chronic symptoms started to report that their sense of smell was improving or that they weren’t as fatigued. “For some of them it was short lived. But for a chunk, it actually persisted — they went ahead, got their second shot out, and are saying, wow, they really feel like there’s light at the end of the tunnel,” Griffin says.

There are plausible biological reasons vaccination could help people with long COVID, says Akiko Iwasaki, an immunologist at Yale University. Scientists still don’t know for sure why some people have chronic symptoms, but one theory is that the virus or fragments of the virus stick around in their body. They’re not contagious, but the leftovers continue to irritate the immune system. Vaccination could clear those out. “Potentially, those remnants are removed because you’re generating a lot of antibodies,” Iwasaki told The Verge.


There are a number of reports to this effect. That’s very encouraging.
unique link to this extract



CDC’s draft guidelines for vaccinated Americans call for small steps toward normal life • POLITICO

Erin Banco:


The CDC guidance, which could be released as early as Thursday, will include recommendations that Americans limit their social interactions to small gatherings in the home with other fully vaccinated individuals, wear masks in public and adhere to other public-health measures such as social distancing for the foreseeable future.
But the agency’s advice is likely to disappoint many who hoped the increasing pace of inoculations would allow some common restrictions to be relaxed immediately for vaccinated people.

The document will include a series of scenarios for Americans to consider, including where they socialize, with whom they can socialize with and what to consider when making plans. It will also include a section on travel.

Dr. Anthony Fauci, President Joe Biden’s chief medical officer, as well as CDC Director Dr. Rochelle Walensky previewed the guidance at a press conference with reporters Monday. Fauci said while the guidelines were still being finalized, “doubly vaccinated” Americans could gather within the home safely.

“I use the example of a daughter coming in from out of town who is doubly vaccinated, and a husband and wife doubly vaccinated, and maybe a next-door neighbor who you know are doubly vaccinated,” Fauci said. “Small gatherings in the home of people, I think you can clearly feel that the risk — the relative risk is so low that you would not have to wear a mask, that you could have a good social gathering within the home.”


Hardly thrilling. A whole lot of doubly vaccinated people, and you finally think they might be able to leave their masks off? That doesn’t show that much confidence in the vaccine.
unique link to this extract



Colleges that require coronavirus screening tech struggle to say whether it works • The New York Times

Natasha Singer and Kellen Browning:


Before the University of Idaho welcomed students back to campus last fall, it made a big bet on new virus-screening technology.

The university spent $90,000 installing temperature-scanning stations, which look like airport metal detectors, in front of its dining and athletic facilities in Moscow, Idaho. When the system clocks a student walking through with an unusually high temperature, the student is asked to leave and go get tested for Covid-19.

But so far the fever scanners, which detect skin temperature, have caught fewer than 10 people out of the 9,000 students living on or near campus. Even then, university administrators could not say whether the technology had been effective because they have not tracked students flagged with fevers to see if they went on to get tested for the virus.

The University of Idaho is one of hundreds of colleges and universities that adopted fever scanners, symptom checkers, wearable heart-rate monitors and other new Covid-screening technologies this school year. Such tools often cost less than a more validated health intervention: frequent virus testing of all students. They also help colleges showcase their pandemic safety efforts.

But the struggle at many colleges to keep the virus at bay has raised questions about the usefulness of the technologies. A New York Times effort has recorded more than 530,000 virus cases on campuses since the start of the pandemic.


More Covid theatre: you can have the virus but not have a temperature, and possibly be spreading it. Not linking the positive temperature results to the virus outcome is just mindboggling, though. (Thanks G for the link.)
unique link to this extract



The ‘LitterCam’ that’s watching you • BBC News

Justin Rowlatt:


CCTV cameras will soon have a new target – litter louts.

AI software can now match footage of motorists throwing rubbish to their car’s number plate and issue an automatic fine of £90.

The first trial of the potentially controversial new system will begin in Maidstone in Kent in a few weeks with other councils expected to follow.


Can even, they say (and show) detect a cigarette butt being thrown out of a window, which is quite an extreme interpretation of “littering”. Automatic numberplate recognition then records the car details.
unique link to this extract



this app works by typing
new lines create new nodes
indentation creates child nodes
and any text: before a colon+space creates a label
[linking] you can link to nodes using their ID in parentheses
like this: (1)
lines have a default ID of their line-number
but you can also supply a custom ID in brackets
like this: (linking)


Your diversion for today. Creates lovely, possibly silly, possibly useful flowcharts (though not decision trees).
unique link to this extract



Oakland bans the use of combustion engine-powered leaf blowers and string trimmers • City of Oakland



Prohibition on Combustion Engine-Powered Leaf Blowers and String Trimmers Ordinance (OMC 8.64)

Combustion engine-powered leaf blowers and string trimmers are those powered by an internal combustion or rotary engine using gasoline, alcohol, or other liquid or gaseous liquid. These devices pose significant health hazards to both equipment operators and Oakland residents, including the discharge of particulate matter, carbon monoxide, unburned fuel, and ozone. They also contribute to climate change by emitting carbon dioxide, a greenhouse gas, and generate significant noise pollution, a paramount concern for Oakland residents.


So it’s the noise more than the greenhouse gases? Oakland is, officially, a city with about 433,000 residents. That’s probably quite a lot of leaf blowers and strimmers, which (because they use two-stroke engines) will pollute as badly as a large car does.

So this ordinance, passed in January, might seem silly. But it will have real impact.
unique link to this extract


Errata, corrigenda and ai no corrida: Fosco Marotto, the CTO of Gab, posted on Hacker News explaining that his “simple” coding error that led to Gab getting comprehensively hacked was in fact a very complicated coding error where he’d been completely aware of the risks of SQLi but had thought that all the data that would be presented to the query would be sanitised (as it’s known). Which goes to show that SQLi, like rust, never sleeps. (Thanks Seth for the link.) 

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.