• 2 Posts
  • 20 Comments
Joined 3 months ago
Cake day: April 7th, 2025
  • ravenaspiring@sh.itjust.workstoTechnology@lemmy.ml•3 Teens Almost Got Away With Murder. Then Police Found Their Google Searches
    101·
    10 hours ago

    Some relevant quotes to summarize:

    But the videos weren’t clear enough to identify the exact make or model of the dark four-door sedan. The detectives quickly obtained what are known as tower dump warrants, which required the major phone networks to provide the numbers of all cellular devices in the vicinity of 5312 Truckee during the arson. And they slung a series of so-called geofence warrants at Google, asking the company to identify all devices within a defined area just before the fire. (At the time, Google collected and retained location data if someone had an Android device or any Google applications on their cell phone.)

    …

    There were 1,471 devices registered to T-Mobile within a mile of the house when it ignited. Using software that visualizes how long it takes a signal to bounce from a cell tower to a phone and back again, Sonnendecker narrowed the list down to the 100 devices nearest to the house. One evening toward the end of August, detectives roamed the area around 5312 Truckee with a cell-phone-tower simulator that captured the IDs of all devices within range. That night, there were 723. Sonnendecker cross-referenced these with the 100 from earlier, eliminating the 67 that showed up on both lists and likely belonged to neighborhood residents who could be ruled out. That left 33 T-Mobile subscribers whose presence in Green Valley Ranch in the early hours of August 5 couldn’t easily be explained.

    …

    That’s when another detective wondered if the perpetrators had Googled the address before heading there. Perhaps Google had a record of that search?

    … birth dates, and physical addresses for all users who’d searched variations of 5312 Truckee Street in the 15 days before the fire.

    Google denied the request. According to court documents, the company uses a staged process when responding to reverse keyword warrants to protect user privacy: First, it provides an anonymized list of matching searches, and if law enforcement concludes that any of those results are relevant, Google will identify the users’ IP addresses if prompted by the warrant to do so. DPD’s warrant had gone too far in asking for protected user information right away, and it took another failed warrant 20 days later and two calls with Google’s outside legal counsel before the detectives came up with language the search giant would accept.

    Finally, the day before Thanksgiving 2020, Sonnendecker received a list of 61 devices and associated IP addresses that had searched for the house in the weeks before the fire. Five of those IP addresses were in Colorado, and three of them had searched for the Truckee Street house multiple times, including for details of its interior. “It was like the heavens opened up,” says Baker.

    In early December, DPD served another warrant to Google for those five users’ subscriber information, including their names and email addresses. One turned out to be a relative of the Diols; another belonged to a delivery service. But there was one surname they recognized—a name that also appeared on the list of 33 T-Mobile subscribers they’d identified earlier in the investigation as being in the vicinity of the fire. Bui.

    …

    Seymour’s defense argued that, in asking Google to comb through billions of users’ private search history, investigators had cast an unconstitutional “digital dragnet.” It was, they said, the equivalent of police ransacking every home in America. The Fourth Amendment required police to show probable cause for suspecting an individual before getting a warrant to search their information. In this case, police had no reason to suspect Seymour before seeing the warrant’s results. But the judge sided with law enforcement. He likened the search to looking for a needle in a haystack: “The fact that the haystack may be big, the fact that the haystack may have a lot of misinformation in it doesn’t mean that a targeted search in that haystack somehow implicates overbreadth,” he said

    …

    After a five-month wait that Sandoval remembers as “gut-wrenching,” the court finally ruled in October 2023. In a majority verdict, four judges decided the reverse keyword search warrant was legal—potentially opening the door to wider use in Colorado and beyond. The judges argued that the narrow search parameters and the performance of the search by a computer rather than a human minimized any invasion of privacy. But they also agreed the warrant lacked individualized probable cause—the police had no reason to suspect Seymour before they accessed his search history—rendering it “constitutionally defective.”

    Because of the ruling’s ambiguity, some agencies remain leery. The ATF’s Denver office says it would only consider using a keyword warrant again if the search terms could be sufficiently narrowed, like in this case: to an address that few would have reason to search and a highly delimited time period. The crime would also have to be serious enough to justify the level of scrutiny that would follow, the ATF says.

    …

    Meanwhile, another case—in which a keyword-search warrant was used to identify a serial rapist—is now before the Pennsylvania Supreme Court. If the warrant is upheld, as it was in Colorado, their use could accelerate nationwide. “Keyword warrants are dangerous tools tailor-made for political repression,” says Crocker. It’s easy to envision Immigrations and Customs Enforcement requesting a list of everyone who searched “immigration lawyer” in a given area, for instance.

  • ravenaspiring@sh.itjust.workstoWorld News@lemmy.ml•Iran wants to buy a new air force from China
    7·
    4 days ago

    Immediately following American air strikes, Iranian military officials jetted to Beijing to negotiate the purchase of J-10C fighter jets and AWACS from China.

    Pakistan’s air force is equipped with Chinese-built fighters and missiles, and in a recent air battle against Indian forces shot down several top-line Western-built fighter jets and drones.

    Chinese defense firms build equipment that is comparable or superior in quality, at considerable cost savings head-to-head against gear from NATO countries or Russia.

    What’s more, other militaries are struggling to build weapons at all, given that supply chains for the most advanced munitions and systems run through China for rare earth metals, and the other BRICS countries for raw materials.

    If China agrees to supply Iran, it will remake the military and diplomatic landscape of the Middle East, for decades to come

    Interesting times… /s

  • ravenaspiring@sh.itjust.workstoUnited States | News & Politics@midwest.social•Several Polls Show Trump’s Approval Rating Is Underwater by Double Digits
    3·
    12 days ago

    While the message is not entirely off, it’s worth realizing that polling approval numbers are all over the place. While Nate Silver might not peg it, be does at leas provide more polls which shows this:

    As he says:

    Now these differences aren’t too surprising. It’s normal for individual polls to disagree because of sampling error. But variation in how polls are conducted (whether they interview adults or registered voters, the variables they weight on, etc.) can make those differences even larger. For example, Trafalgar, InsiderAdvantage, and RMG all have Republican house effects while Ipsos tends to have a strong Democratic house effect. …

    Inevitably, there’s a lot of disagreement from survey to survey, not just because of statistical variation but because pollsters have long had trouble pegging down Trump’s popularity — and often underestimated it.

    Which is to say, polling is still more of an art.

    This is not to dampen the delight too much, but reality is much more complicated than a poll.

  • ravenaspiring@sh.itjust.workstoPrivacy@lemmy.ml•Meta: Coming after your data harder than ever
    52·
    17 days ago

    This android only.

    From the article:

    Meta managed to do this even when:

    • You aren’t using the app (but have a session open in the background).

    • You haven’t logged into your account in the browser.

    • You’re browsing in incognito mode.

    • You’re using a VPN.

    • You delete cookies at the end of every session.

    The captured data includes:

    • Complete browsing history with specific URLs

    • Products added to cart and purchases made

    • Registrations on websites and completed forms

    • Temporal behavioral patterns across websites and apps

    • Direct linking to real identities on social networks

    You’re not affected if (and only if)

    • You access Facebook and Instagram via the web, without having the apps installed on your phone

    • You browse on desktop computers or use iOS (iPhones)

    • You always used the Brave browser or the DuckDuckGo search engine on mobile

  • ravenaspiring@sh.itjust.workstopolitics @lemmy.world•How Amy Coney Barrett Is Confounding the Right and the Left
    13·
    23 days ago

    She has become the Republican-appointed justice most likely to be in the majority in decisions that reach a liberal outcome, according to a new analysis of her record prepared for The New York Times. Her influence — measured by how often she is on the winning side — is rising. Along with the chief justice, a frequent voting partner, Justice Barrett could be one of the few people in the country to check the actions of the president.

    Overall, her assumption of the seat once held by Justice Ruth Bader Ginsburg has moved the court’s outcomes dramatically to the right and locked in conservative victories on gun rights, affirmative action and the power of federal agencies. But in Trump-related disputes, she is the member of the supermajority who has sided with him the least.

    So not what they wanted, but not RBG. Still to much authoritarianism for me from the SCOTUS.

  • ravenaspiring@sh.itjust.workstopolitics @lemmy.world•What the 3.5% Rule Tells Us About Protest Success
    162·
    23 days ago

    Fascinating idea and I look forward to reading the book. As someone who has never seen protests be that effective as compared to other constituency pressure mechanisms, it’s an interesting counter point.

    The OP’s article indicates 3.5% of the population, which for the US at the moment would be around 340 million. 3.5% would be 11.9 million people.

    Rough guesses are that the protest saw about 4-6 million people out yesterday.

    I’m particularly curious about the paper’s coalition building concepts about tying immigration to other value such as worker rights, private sector interests such as agriculture, racial justice, etc.

    Beyond this I wonder if the analysis from ten years ago takes into account the technological isolation, manipulation, and echo chambering of modern politics. I would venture to guess that the 3.5% might need to be higher in a population that doesn’t listen to ‘untrusted opinions’.

  • ravenaspiring@sh.itjust.workstoPrivacy Guides@lemmy.one•*Permanently Deleted*English
    3·
    2 months ago

    They should have said, but I guessing they didn’t because it’s easy and indicates it at the checkpoint. You tell the agent you are declining the biometric identification. Do it as soon as you hand the agent your ID, don’t step in front of the machine, then follow instruction (probably step alongside so they have a clear view of your face).