Topps Blog

Monday, 25 October 2021

FTC Study Confirms ISPs Collect a Scary Amount of Your Personal Data

Frustration with Big Tech is reaching a crescendo, and we could be headed for legislative efforts to reduce the influence of companies like Google, Amazon, and Facebook, but another technology heavyweight has managed to fly under the radar: your Internet Service Provider (ISP). A new Federal Trade Commission (FTC) report examines what ISPs are doing with your data, and it’s not good news. Even when they promise not to sell your information, it often ends up in the hands of advertisers and other third parties.

An internet stalwart like Google could have a vast amount of data on you — it’s the price you pay for using Google’s freely available services. That’s the catch: you have to use Google’s services before it has your data, and it doesn’t get everything. The same is true of Facebook, Twitter, and everyone else. If you don’t use the service, they have limited data on your online activities. Your ISP, on the other hand, is the gateway to the outside world, and it can see everything you do.

The FTC report (PDF) analyzed the practices of six different ISPs and their advertising subsidiaries. While the report doesn’t name names, it would be safe to assume the misadventures laid out in the document are standard across the industry. For example, several of the ISPs included in the analysis promise not to share data with third parties, but the FTC found they still do so with the help of shady business deals. The agency says that while ISPs often claim they only retain user data for “business purposes,” this broad definition allows them to engage in these practices with impunity.

Information exchanged by ISPs in the study.

Having a tap on the data leaving your modem gives ISPs deep insight into you as a person. Not only is your online browsing history recorded, but so is your search data, DNS records, financial records, video viewing stats, and even what smart home devices are active and when. The FTC notes that the business justification of keeping much of this data is shaky at best, and even when ISPs provide opt-outs for this sort of data collection, the FTC says these “illusory choices” are usually ignored by the ISP.

This all sounds like it should be illegal, right? Or at least in violation of some regulatory rules. The FTC did try to make rules preventing these excesses in 2017, which would have forced ISPs to get opt-in consent to collect data such as your precise location, browsing history, and app usage. However, the GOP-controlled Congress used the Congressional Review Act to block it.

The FTC voted unanimously (4-0) to release this report to the public, but it can’t do much about the problem when Congress has shown it will do whatever it takes to protect the broadband industry. Perhaps the best way to prevent this right now is to route your home broadband traffic through a Virtual Private Network (VPN). However, you will then have to trust the VPN more than your ISP.

Now Read:

Calling Citizen Scientists: You Can Help Find Exoplanets From Your Couch

If you’ve ever wanted to help find an exoplanet, this is your chance. There’s a new open research initiative that invites participants to help identify exoplanets using simple visual pattern matching — and it’s open to anybody who wants to help. You don’t even have to own a telescope.

The project is an open, distributed search for exoplanets, in cooperation with a planet-hunting telescope array called the Next-Generation Transit Survey (NGTS). You can be a seasoned pro-am astronomer with a gigantic Dobsonian on your patio, or you might never have tried to engage in one of these projects before. It doesn’t matter your level of experience; you’re invited.

What they need is help identifying the dip in a star’s brightness when a planet transits in front of it. To do this, participants visually compare images to see whether there is a characteristic shape to the “light curve” of a star’s brightness over time. If the pattern is recognized, it’s tagged for further analysis.

The stars identified as having transiting planets are rare to begin with, because these planets’ orbits have to be precisely edge-on toward us. As a transiting planet swoops around between us and its star, the speed, shape, and eccentricity of the planet’s orbit all impact the rate at which it occludes the parent star’s light. In a sense, transits are like an eclipse, only the blip in a star’s brightness is much smaller because stars are so large.

So far, transit hunting is our most successful way of finding planets outside our solar system. You may remember that NGTS isn’t the only major array using the transit method. TESS also searches for transiting planets, much like Kepler.

The NGTS is an array of twelve 20cm telescopes. Consumers can buy telescopes of that size on Amazon.

Planets orbiting their stars should be as regular as clockwork — but there are things that can influence a star’s brightness as seen by our telescopes.

There’s also the problem of sunspots. Our sun has a whole sunspot cycle, with periods of increased and decreased sunspot activity. Because it oscillates between those two states, an observer at a high or low point in the cycle might not get a representative sample of our sun’s activity. We’ve seen this happen on other stars too. For example, the case of the asymmetrical dimming and subsequent brightening of Betelgeuse from October 2019 through February 2020. A mammoth gas expulsion that dimmed the star and disturbed a relatively large percentage of the star’s surface was determined to be the reason for the weird visuals we saw.

One of the more exotic light curves identified in the NGTS exoplanet hunting project. This shows the oscillations of a variable star with a transiting exoplanet. Credit: NGTS Planet Hunters

Our own atmosphere is another problem. Like other earthly telescopes, NGTS has to deal with the same atmosphere that makes stars twinkle to the naked eye. This is why the twelve telescopes that make up the NGTS array are installed at the ESO’s Paranal observatory. It sits in its aerie in the Chilean Andes, next to the Atacama desert. It’s so dry, and it lies at such an altitude, that it’s as close to space as you can get without actually putting the whole observatory in orbit.

The imaging frequency of NGTS is slower than TESS or Kepler. There’s also intermittent downtime. That’s important because participants are asked to decide two things: whether a given dip in a brightness curve is more like a U, a V, or a flat bottom, and whether a pattern of dips is regular enough to be an orbital transit. Fewer data points makes it tougher to get an accurate sense of a trend. If the image capture and the transit interval were timed just right, the camera could miss a transit every time. This is one place where machines just haven’t quite caught up with humans.

What’s your favorite amateur/citizen/distributed science effort? Did you do SETI or Folding@home? Are you participating in the NGTS project? Let us know in the comments!

Now Read:

Oculus Gives Discontinued Oculus Go a New Lease on Life with Unlocked Software

The release of the Oculus Go in 2018 was an important step for Facebook’s Virtual Reality division, but the product was short-lived. After being made obsolete by the Oculus Quest, the company discontinued the Go in 2020. Still, Facebook sold millions of the standalone headsets, and now they’re even more capable thanks to an unlocked software update. That means anyone with an Oculus Go will be able to tinker, mod, and develop software for them for years to come.

The Oculus Go was the company’s first standalone headset, and it came at a pivotal time. Back then, Oculus was focused on its desktop VR Rift system and a partnership with Samsung for mobile VR. However, Samsung decided to end support for its Gear VR feature. The Go offered a similar experience with access to the same mobile-optimized software that didn’t need to deal with the battery and heat dissipation limitations of a smartphone.

Eventually, the attractive $199 asking price led to about two million sales. Many of those devices are still floating around, and it’s only a matter of time before Oculus completely forgets about it. According to Oculus CTO (and id Software founder) John Carmack, he’s been pushing for unlocked software for years. When a company no longer has a financial interest in keeping a product operational, Carmack believes it should be opened up for users to maintain.

The update, which rolled out to the Oculus Go on October 21st, grants full root access to the Android-based device. Carmack says that even if someone found a shrink-wrapped Go in 20 years, they would be able to start it up and load the most recent software many years after Facebook shut down the official OTA servers. This might even make existing GO headsets a hot commodity. Units are still available on eBay for as little as $50, and they all have the potential to foster the development of cool homebrew VR software.

This is a great policy and not the sort of thing we’re used to seeing from a Facebook brand. I can only hope the same generosity applies to newer Oculus devices in the future—an unlocked OS for the Quest or Quest 2 would be fantastic. Other technology firms could also take a page from the Oculus playbook here. Many products end up non-functional after companies deactivate their aging online services. Game console makers are particularly bad about this, often sunsetting stores and gaming platforms on last-gen hardware while people are still using them. It’s a very user-hostile policy. For once, can’t everyone just be more like Oculus?

Now Read:

Analysis Shows Amazon’s Alexa Collects More Data Than Any Other Smart Assistant

(Photo: Jan Antonin Kolar/Unsplash) When it comes to smart assistants and data privacy, it appears Amazon’s Alexa is your worst bet. A new analysis conducted by Reviews.org has revealed that out of the smart assistants you could have on your phone or in your home, Alexa collects the most user data—including everything from personal identifiable information (PII) to shopping habits.

Reviews.org compared Amazon’s Alexa, Google’s Assistant, Apple’s Siri, Samsung’s Bixby, and Microsoft’s Cortana on the basis of which types of data each collected from their users. While all five smart assistants collected the user’s name, time zone, phone number, location, and IP address, Alexa was found to gather more user data than the others: the device also takes note of the user’s address, age, voice characteristics, payment information, and personal interests. It gathers the last piece from the user’s profile, along with a “personal description” stored in said profile. And this is only the user’s personal data—other information, like contacts, media, communication records, app use, Wi-Fi network details, and ISP details are also stored by Alexa, but not necessarily by all four other smart assistants.

Personal data collected by the five major smart assistants. (Image: Reviews.org)

By now most of us know that when we purchase a new device, download a new app, or create a new online profile, we’re opening the door to complicated privacy concerns. Many of us have begrudgingly come to a sort of unspoken compromise: limited data collection is something we’re willing to put up with in order to enjoy the internet. But a line has to be drawn somewhere, and while some people may be okay with in-home devices taking careful notes about how their voice sounds and what they’re buying, others aren’t.

A survey conducted in tandem with the analysis by Reviews.org shows that a majority of smart assistant users are concerned about having their data collected or owning a device that’s “always listening” in the background. Though there are (supposedly) ways to turn off the mic on a smart assistant, 38 percent of users reported being unsuccessful in doing so (compared to the 45 percent of users surveyed who tried).

Of course, the easiest way to keep a smart assistant from collecting your data is to avoid buying one in the first place—but if you must, Alexa might not be your best option.

Now Read:

Friday, 22 October 2021

Twitter’s Internal Research Confirms Its Algorithm Favors Right-Wing Voices

In the last 15 years, Twitter has gone from an SMS-based messaging platform to a social network with hundreds of millions of users. In recent years, conservative politicians and activists have railed against “cancel culture,” claiming Twitter and other social media outlets are biased against them—hence, the existence of right-wing alternatives like Gab, Parler, and Trump’s new Truth Social. However, Twitter has just published some internal research that suggests the opposite. According to its research, its algorithm actually favors right-wing voices.

You can read the full report, which is available for download directly from Twitter. The 27-page PDF leverages the expertise of computer science researchers from the University of Cambridge, University College London, and the University of California – Berkeley. It focuses on the Home timeline, which is the algorithmically driven feed that everyone sees by default in Twitter.

The researchers analyzed millions of tweets that appeared in users’ Home feeds between April 1 and August 15, 2020. The tweets identified for use in the study were from elected officials in seven countries. They used public data like official websites and political registration to determine who counted as a conservative voice. Then, the researchers crunched the numbers to compare the pure reverse chronological Home layout to the default “Top Tweets.” There are several key findings, some more concerning than others. For example, it should come as no surprise that during 2020, the algorithm amplified political speech in general. However, just because one political account ranked higher, that didn’t mean everyone in the same political group would see the same boost.

On the flip side, Twitter confirms that conservative accounts are being favored by the algorithm. This is true in six out of the seven countries studied, the lone exception being Germany, which is particularly sensitive about right-wing speech for obvious reasons. In addition to conservative politicians getting top billing, the news outlets they promote are also seeing wider promotion in the algorithm. Twitter’s ongoing quest to hobble third-party apps also ensures most of its user base will see the algorithm-based feed, which they might not have in the past.

Currently, Twitter doesn’t know how to fix this, but it seems like it wants to. That’s more than we can say for Facebook, which is again in hot water after a former employee released a trove of materials that shows how the network tunes its algorithm to increase engagement by getting people outraged. Twitter says it is finalizing a partnership that will allow it to release large datasets to researchers without impacting user privacy. With additional research, Twitter might be able to find the trigger for these inequities. In the meantime, you can always manually switch from Top Tweets to chronological order.

Now Read:

US Army Corps of Engineers Uses Azure for Storm Modeling

(Photo: Brian McGowan/Unsplash) The U.S. Army Corps of Engineers and Louisiana’s Coastal Protection and Restoration Authority are teaming up with Microsoft to better prepare the state for hurricanes. The agencies are using Azure Government, a Microsoft computing and analytics tool reserved specifically for the government, to model synthetic storms and determine the best ways to prevent the need for future evacuation and reconstruction.

After the U.S. Army Corps Engineer Research and Development Center (ERDC) kicked off the pilot project under the Department of Defense’s High Performance Computing Modernization Program (HPCMP), Microsoft awarded the ERDC with an AI for Earth grant, which aims to “put Microsoft Cloud and AI tools in the hands of those working to solve global environmental challenges.” The Louisiana authority and the Army Corps will run CSTORM-MS, a high-capacity storm modeling system, on Azure to explore the use of artificial intelligence and machine learning in garnering storm insights. They’ll start by studying how seawater could surge up the Mississippi River during a major storm and affect coastal communities.

The Mississippi River, a major economic driver for Louisiana, is also a flood risk during hurricanes and other major storms. (Photo: Justin Wilkens/Unsplash)

Following such catastrophic disasters as Hurricanes Ida and Katrina (as well as the growing urgency of climate change), it’s imperative that the government find ways to mitigate the effects of tropical storms. Using Azure, the Army Corps can decide which flood-risk infrastructure to prioritize, as well as where to place it. Seawalls and levees are critical to temper hurricanes’ impact on nearby communities, but precisely where they’re constructed makes a world of difference. Putting a levee in one spot can lead to flooding in another, for instance, and room for error is slim. “If you’re building a levee that goes 10, 20 miles or so and you’ve got to have an error tolerance on how high that levee is, you’re talking millions of dollars if you’ve got to build it higher,” ERDC research mathematician Chris Massey told Microsoft.

The agencies have already mimicked over 2,300 computerized storms prior to bringing Azure Government onboard. Resources have historically been scarce; CSTORM’s heavy lifting requires the team to use Department of Defense supercomputers, which are in high demand by researchers involved in COVID-19 mapping and military analyses. Azure will allow the ERDC to circumvent potential delays caused by this demand and be able to respond efficiently during an active storm.

Now Read:

AT&T’s Free Batman Cloud Gaming Points to a Different Future for Google’s Stadia

Google unveiled Stadia in 2019 with the promise of a platform that would rival the traditional gaming consoles. Barely a year later, it is begun backing away from that promise. Google has already closed its game studio, and has previously said it would prefer to just license Stadia technology to other companies. Now, it appears the first product of that approach has arrived. AT&T Wireless customers can stream Batman: Arkham Knight for free, but not via the Stadia app or website.

For anyone living under a rock who has missed Google’s bumbling Stadia journey, the service is similar to GeForce Now or Xbox Game Pass (née xCloud). A server renders the game and streams video to your device. Your control inputs go back up to the server, allowing you to play the game. Latency under ideal conditions can be close enough to local play that you won’t notice in most games. Of course, conditions are often far from ideal. Still, some of these services can work surprisingly well, depending on where you live, the strength of your internet connection, etc.

Google seems to have realized too late that building the powerful infrastructure to run Stadia was the easy part. Designing AAA games is hard, expensive, and takes years. Google wasn’t into that, so Stadia Games & Entertainment was canned. At the same time, it said that licensing the tech was “the best path to building Stadia into a long-term, sustainable business.” The AT&T deal might be the first of many in this vein, allowing a company to provide game streaming without developing its own system. Why AT&T chose a Batman game from 2015 is unclear, but it does own Warner Media, which in turn owns DC comics.

Mortal Kombat 11 was one of the big AAA launch titles for Stadia, but it hasn’t gotten many since then.

Everyone can play a limited selection of Stadia content for free, but the Batman game is only for AT&T subscribers. Simply head to the landing page, input your account info, and you can begin playing the game instantly. There’s no Stadia branding anywhere on the AT&T landing page, so most people will never know it’s using Google’s tech. Ars Technica has confirmed with AT&T that the free game stream is powered by Stadia. However, you might notice something weird if you go to the Stadia site: there’s no listing for Batman: Arkham Knight.

So, Google ported Arkham Knight to Stadia, but it’s not being made available generally, which is a bummer. Perhaps it will come to the Stadia store at a later date, but there’s no guarantee. We might see more of these exclusive streaming offers, but that strikes me as the wrong strategy if Google cares about Stadia as a consumer product—and maybe it doesn’t! The current game offerings are weak, and making new games for Stadia that won’t be beefing up the store’s content almost guarantees things will just get worse.

Now Read: