AramZS. he/him. Privacy Engineer. Tech for journalism. Ad Tech. Prev: Fullstack, strategy, econ/game journo, storytelling, altac. Views are only my own
The numbers are all fking fake, the metrics are bullshit, the agencies responsible for enforcing good practices are knowing bullshiters enforcing and profiting off all the fake numbers and none of the models make sense at scale of actual human users.
I do think that people are really underestimating the risk in a post-Roe world that ad tech poses to women. It isn't just Law Enforcement has an easier time accessing your phone than your home and it isn't just data brokers. States now have private-citizen bounties, remember?
People are writing long think-pieces trying to figure out wtf happened this decade. I can explain the whole decade in two photos.
On the left: 2010, on the right: 2019
As folks in some states are preparing for what to do next, please remember that the underlying technology of ad tech has been used to prosecute both doctors and those who receive abortions.
I do think it's wild that over the last 365 days almost every non-profit newsroom (including many NPRs and big names everywhere) have lost major funders and each time it's treated as an individual story and no one asks if there is an underlying issue.
🟡 SCOOP: The Intercept is nearly out of money and facing a bitter civil war, with multiple feuding factions battling for power and two star journalists trying to take control,
@maxwelltani
reports.
An interesting little fact about Vice's content: a full site archive, including saving outbound links, was performed by the volunteer Archive Team last year and it took ~six months to capture all the Vice content across all the languages they publish in. They've published a lot!
Wow, now that the NYPD has taken over parking for most of a block in my neighborhood I'm starting to see just how many cops have extremely expensive BMWs.
The thing that most people don't seem to understand is that surveillance capitalism means that the systems of ad tech can be very good at tracking you individually, great for use by cops, pretty nifty for propoganda, and still terrible for satisfying standard marketing outcomes.
The problem isn't just that the internet is full of fakery and bullshit and bad numbers and malfunctioning metrics and bullshitters and fraudsters. The problem is that all the fake shit is layered on top of other fake shit and it just COMPOUNDS itself...
OMNY is intended by the MTA to phase out Metrocard. If it does, that means a private company will have detailed travel records, linked to cellphone and credit card IDs, about the millions of New Yorkers who use subways or buses in our city...
Hello to the maybe half of my new followers who are humans. If you came here because of my tweet about fake traffic and ad tech, buckle in because that's exactly the topics you're going to get tweets about from this account, day and night.
OMNY owning the machines that give out metrocards mean they will doubtlessly be applying the extremely creepy tap to pay methodology that can link your phone with your travel, only now it will also link your phone with a travel card with a persistent ID.
You cannot apply Tech Startup Economics to a Journalism company and expect it to operate ethically. All VC-based startups are operating on a set of incentives which are already unethical, but doubly so for journalists
OMNY is not a project of the city. It is a system operated by the private organization Cubic Transportation Systems. Cubic has quietly taken over the monetization process of the majority of major US and western EU cities -
I'm well acquainted with video, building video code, and building video metrics & this doesn't surprise me in the least. At previous gigs, I spent years calling bullshit on video as a strategy internally. The numbers I saw with my own tools never matched what was promised
To understand the consequences of the Metrocard being taken over by Cubic, we don't have to look much further than London's Oyster card, whose privacy problems are well documented. In one year Metro Police made over 3,000 requests to Oyster user records...
The problem is that the scale that VCs are defining and that Mic is trying to chase is a lie. The numbers were never really there. Eventually they were always going to disappear as fraudulent traffic and metrics fell apart. This is still occurring.
If I was a piece of shit and good at ad targeting or just had a bunch of buyer accounts I could say... arbitrage my way to a $10k bounty in Texas at scale. This doesn't even get into how cops could use targeting data without ever writing a warrant.
So this represents a fun set of assumptions.
1. You do all your reading based on things you click on Twitter
2. Twitter tracks every single article you click on and records the full URL against your profile.
3. Retweeting unread links is very common
Sharing an article can spark conversation, so you may want to read it before you Tweet it.
To help promote informed discussion, we're testing a new prompt on Android –– when you Retweet an article that you haven't opened on Twitter, we may ask if you'd like to open it first.
Now that Facebook is pivoting to video again, I thought it might be useful to go over a rough timeline of the on-again-off-again pivots to video over the last 5+ years...
It really is remarkable how quickly AI is falling into the ad tech model of 'you shouldn't ethically do this even if it worked, but also it doesn't work anything like as promised and that's part of the problem'.
“A security firm that sells AI weapons scanners to schools is facing fresh questions about its technology after a student was attacked with a knife that the $3.7m system failed to detect.”
@JamesClayton5
This could have serious consequences if, like many other big data systems, it becomes accessible without warrants. Consequences for immigrants could be terrible
The ad tech ecosystem doesn't need to be pruned. It needs to be burned to the ground. Until that happens everything that's wrong with the internet will continue to just get worse, because ad tech creates the *incentive* to make it worse.
So I ran an export of my data from Twitter and guess what! There's a new file that wasn't there in 2018 'user-link-clicks.js' which logs links I clicked on, what tweets they were in, the url those links resolve to & the time stamp of when I clicked, starting 05-13-2020.
I cannot discourage tap-to-pay enough. Your credit card being with these people is awful. The NFC signal that's transmitted with the tap, the doubtless present Bluetooth beacons, the ability to combine that with other data, it's really bad in terms of identifying you...
I'm well acquainted with video, building video code, and building video metrics & this doesn't surprise me in the least. At previous gigs, I spent years calling bullshit on video as a strategy internally. The numbers I saw with my own tools never matched what was promised
Turns out people *don't* prefer to watch a 7 minute video of a guy reading a 30-second news story.
Good one, Facebook! Thanks for [checks notes] irreversibly destroying the media landscape and [checks notes] fracturing democracy the world over.
The terms OMNY offers are very unclear on privacy promise and if you think that the NYPD can't get at your data without a warrant... history says otherwise.
When asked NYC Transit officials do not have answers about how our privacy will be protected.
Worth noting: comScore numbers are a joke. Everyone in the biz side and plenty on the editorial side know it. But we keep using them anyway and I guess they are sorta interesting in terms of comparison.
I'm going to rule this one for BuzzFeed I guess.
An activist or LEO ad buyer who *suspects* a person of having an abortion could weave together targeting, user IDs, app & web data to build a supporting case. They could buy data pretty cheap OR it could mean attempting to target a user & de-anonymize.
And if you don't tap to pay with your phone for a Metrocard, you are likely paying with plastic. If you are, then your card management company (VISA, etc...) will have your data and there will be a major data point to join OMNY data with Metrocard data ...
"On a traffic flow sample of over 400m installs over 17 days, we estimated that $1.7m worth of installs were being paid to fraudsters faking installs"
$1.7 million *just* in fake installs, over the course of less than 3 weeks!
So that answers my question:
yes, instead of storing metadata about links you click now Twitter stores a log that explicitly contains everything you've ever clicked, the tweet you got a link from, & precise timestamp of that click (great for deanonymizing cross site activity!)
A >500 mil downloads app is listening to when you download other apps and firing off background ad requests in order to fake ad calls from newly installed apps.
Autoplay video inserted into a feed by an advertiser because Twitter lacks preroll inventory is something else. It means Twitter is almost - but not quite - lying about viewer counts and the play counts in this video are zero indicator of its success.
Remember... the New York City subway system is not providing you a substantially different service thru OMNY. NONE of this collection or monitoring is *necessary*. The service you get is fundamentally identical to the one you could access using entirely anonymous physical tokens.
I can think of few things more invasive than your tax software becoming a data broker that sells everything it knows about you to allow you to be targeted with ads.
Exclusive: Intuit is starting a media network. Advertisers will be able to target customers of the small business accounting software firm QuickBooks across the web.
Another example of how anything can become an ad network
A friend just discovered a fake Facebook account pretending to be him with his photo has been trying to sell people sunglasses and other crap for >10 years.
There is a *reason* that privacy rights are so tied up with the pro-choice cause and questions of abortion. And we should expect that anything that threatens privacy will intrinsically threaten your right to choose and your bodily autonomy.
As folks in some states are preparing for what to do next, please remember that the underlying technology of ad tech has been used to prosecute both doctors and those who receive abortions.
I read this from
@BenMullin
about the new digital video audience and got a flashback to the pitch for Fusion. All these amazing digital video numbers the IAB likes to report yet no one seems to make money off of digital video.
The internet is full of fake bullshit, run on fake numbers, profitable on fake models, feeding you fake information, supporting fake users. None of this is by accident. Make good choices, build the right technology, tell Congress to start making laws.
In case you were wondering what type of news Buzzfeed considers unprofitable and not worth investing in:
"The buyout is only available to reporters and editors who cover investigations, inequality, politics or science"
And that's the story of how a number of major news sites hosted ads that used illegally stolen images from a site about LITERALLY SHOOTING YOURSELF IN THE FOOT to link to a bullshit fake news website that itself was ripped off for a product that is itself quite dubious.
This is a good time to tell a little story I was told a while back see EverQuote, my favored enemy in the shitty ads world, markets itself to users as a way to get them the best car insurance. But their model makes no sense for that...
The fact that Wells Fargo can continue going as a brand without even a sneaky name change shows how completely Americans have given up any sense of control in regard to government oversight or feeling that they can hold corporations responsible.
Taking a break from tweeting about ad tech to tweet about the attempt by an illegal mob to take over the US capital, an event that is a pretty direct result of ad tech.
I worked with a third party video player that consistently delivered 2x the numbers of my internal metrics. Building metrics isn't hard. There are a ton of metrics providers because it is easy to build analytics and easier to lie with code.
Look my friends, you can embrace the truth with me: App Stores and Apps are bad. It's not just some specific sub-section policy that impacts a single corner of the ecosystem. It's the whole concept, the whole idea, the core practice.
Phone Apps are bad.
Google proposing to resell the rendered product of scraping news sites *back to the news sites* after extracting the maximum value out of it in other contexts and using that profit to further refine it really is the lowest point of their extraction model.
Once you have a de-anonymized ad ID, it opens all sorts of nasty opportunities for anti-choice action. You could leverage segments to find out which you can successfully target the user with.
@swodinsky
@kashhill
I'm not thinking that the app would segment that way, but get the pregnant audience and retarget that segment against abortion keywords is one thing I have in mind. Or delta the pregnant audience with new-mother audience.
Second: if you click on it, the ad goes to a fake news site. It is designed to look like a news site and clearly designed--considering the clear white-background ad--to look as if it was *on* the site you clicked from.
Never forget that if the trajectory of all internet-touching companies is to become ad systems then the truth is that all companies are just masquerading as one thing until they reveal they're surveillance companies.
This is how surveillance capitalism fundamentally works.
ONMY collects:
- Your disability status
- Geolocation information
- Online Activity, with the help of a third party, including:
- Device identifiers
- Cookies
- IP addresses
- "tracking information we or a third party may collect"
...
Oyster's system also categorizes users and assigns them persistent "anonymized" user IDs. Now think about all the digital ad screens in the subway that could make heavy profits mining that data for even more detailed Out Of Home targeting.
Once you either suspect a user OR have geo data about the user (very possible - ) it becomes very easy to find intersections with other data through brokers or active targeting.
Every non profit journalist I know is talking about how no one has money any more and it's wild to me that no one has even tried to explain this phenomenon which sure feels like a systematic issue and not a bunch of random coincidences.
I'm sorry what? Does Vance realize there were two gangs in "Gangs of New York" and one was the anti-Catholic nativist xenophobes and the other gang was the Irish? Are the "ethnic enclaves" JD Vance fears... Irish people? Is JD Vance afraid of the Irish? Weird.
Vance: Has anybody seen the movie "Gangs of New York?." That is what I'm talking about, with these ethnic enclaves in our country, it can lead to higher crime rates.
In case you thought that I was being hyperbolic about the type of information OMNY might collect. Here's their privacy policy's (which you automatically accept when you use their service) list:
- Personal Information including your name, age, photograph, email, DoB
...
Everything is fake, no one knows what is going on, there's not even a mechanism to create trust, much less incentive to use it. The only people in the internet ad tech game you can trust are the ones who this is hurting the most: legit advertisers, legit publishers, users.
This is getting a bunch of activity so I'm going to add, Archive Team is a really cool collective of volunteers who do work on their own to save the history of the internet and if you have a little extra cash, maybe support their efforts!
Don't forget either, the 7000+ ad products that suck money out of the ecosystem as middlemen but don't appear to be producing anything useful except worse user experiences
not a bubble nope not a bubble:
"The Ghostery privacy tool has thousands of trackers in its database, and it's adding about 20 to 30 new trackers per week."
Like, did America's oligarchs suddenly decide they're too rich to bother with the PR boost around funding journalism? Did the money move elsewhere? Are any of the foundations involved even willing to talk about it? I haven't seen a single attempt to answer these questions.
And in a state where there is a bounty on abortions like Texas, the relatively low cost of acquiring this data and running targeting to build your case means that for an unscrupulous individual or anti-choicer this could be a way to arbitrage ads for profit.
So maybe a dragnet is unlikely; but it is easy to see how anti-choice sh*tlord LEOs or anti-choice activists might buy geo data and then use other data or targeting (like ) to try and harvest a whole host of potential targets for prosecution.
Cubic is owned by Veritas Capital
Veritas Capital, among other things, helps run the DHS's biometrics database. Their CCPA statement says they collect and share information collected from "portfolio companies" that includes Identifiers.
Issues include, in no particular order:
- The points system and how it pushes interactions.
- The founder's last company and statements about his opposition to privacy work at Google.
- The policy around prerolled video
- The apparent lack of alt text options...
You can see if you could successfully target a user using a particular apps data that tells you something about if they are currently or actively having regular periods. You could cross reference user data with various geo data sources. You could check to target installed apps.
So *any* system that aims to track you and associate that tracking to information about your body and therefore--to some extent--violate your bodily integrity will inevitably become a tool of those who wish to deny you bodily autonomy & the right to choose
Look folks, we don't have to argue about if period trackers are doing abortion surveillance that can be used by state law enforcement. They are definetly doing some level of surveillance that is neither worthwhile or ethical and will almost certainly be used to do bad sh*t.