The secrets Google spilled in court

“Google’s revenue-sharing deal with Apple was a major part of the trial because Apple is believed to get the bulk of what Google pays out in those agreements. Having a default search placement on Apple devices, which make up roughly half of the smartphone market in the US, is extremely important to Google. We’ve known for years that Google pays Apple for that default placement — this also stops Apple from developing its own search engine — but that’s about it. While Google tried to keep virtually everything about the deal away from the public, we still got a few new details.

In an apparent slip-up, Google’s own witness in the waning days of the trial told us how much of Google’s ad revenue Apple gets: 36 percent for searches done on its Safari browser. The monetary value of that 36 percent is still a mystery. Judge Mehta did not disclose how big Apple’s slice of the $26.3 billion pie is, allowing the DOJ only to say it’s “more than $10 billion.” But the New York Times, citing internal Google sources, put it at $18 billion.”

“We didn’t just find out some of Google’s secrets; a few things about Apple came out, too. Apple’s senior vice president John Giannandrea testified that his company talked to Microsoft about buying Bing in 2018. Apple ultimately decided against it, but not before using the possibility as leverage in its search default negotiations with Google, something Microsoft is still pretty sore about. Apple executive Eddy Cue testified that the company chooses Google to be the default search because it believes Google is the best for its users. But speaking of Bing …”

“Multiple Microsoft executives, including CEO Satya Nadella, testified that Microsoft really, really wanted to make Bing the default search on Apple devices, to the point where it was willing to lose billions of dollars a year for the privilege. Samsung and Verizon, the trial also revealed, essentially refused to even negotiate with Microsoft over changing their search defaults to Bing. Perhaps they were thinking of Mozilla’s experience switching from Google to Yahoo. Mozilla CEO Mitchell Baker testified that Yahoo offered more money and fewer ads, so Mozilla’s Firefox browser switched the default from Google to Yahoo in 2014. Mozilla switched back to Google a few years later, which Baker attributed to Google’s search being better for its users, echoing the point that Google emphasized in its defense.”

https://www.vox.com/technology/2023/11/16/23962967/google-search-antitrust-trial-what-we-learned

The Biden Administration’s Ridiculously Spendy Broadband Promises

“There’s a reason rural and small-town dwellers have less access to the sort of fast internet connections that urban dwellers enjoy: It costs a lot of money to lay fiber-optic cable—”an average cost of $1,000 to $1,250 per residential household passed or $60,000 to $80,000 per mile,” according to Dgtl Infra’s Jonathan Kim. It’s easier to limit and recover costs in densely populated areas where a lot of potential customers live along paved roads than in sparsely settled areas where there’s rough terrain and empty space between each household served. Costs rise dramatically in rural areas.”

“The Alaska Telephone Company, which won a $33 million grant, is planning to run fiber to 211 homes and five businesses at a staggering cost of nearly $204,000 per passing.””

“wireless internet offering 10-50 Mbps and (increasingly) satellite connections such as HughesNet, Viasat, and Starlink are how we connect to the world in my piece of Arizona. It’s a tradeoff you accept if you want open space around you. Well, you accept that tradeoff unless you can rope other people into paying the cost of laying cable.”

“”If you’re spending $50,000 to connect a very remote location, you have to ask yourself, would we be better off spending that same amount of money to connect [more] families?””

https://reason.com/2023/09/08/the-biden-administrations-ridiculously-spendy-broadband-promises/

Meta and Google are blocking links to news in Canada. The US might be next.

“The government legislation that both companies are protesting is called the Online News Act, or C-18. The intention is to give the long-suffering journalism industry a little cash boost, likely at the expense of two companies that are partially responsible for its woes. It accomplishes this by compelling them to pay Canadian news outlets if they host links to their content. (Fenlon’s employer, which is a public broadcaster, officially supports the Online News Act.) That’s why Meta and Google are threatening to remove news links for all Canadian users, permanently, if the law applies to them when it takes effect, likely by the end of this year.”

“The new Canadian law is modeled on a controversial Australian law, the News Media and Digital Platforms Mandatory Bargaining Code, which went into effect in 2021. Google and Meta’s responses to that law were similar threats to pull links, but both companies ended up making payments to some news organizations. The Australian government estimates that news outlets got AU$200 million, although it doesn’t know that for sure — nor does it know how that money was distributed — because the companies were allowed to keep those figures private. Even so, other countries, like Canada, likely assumed they’d get similar results with similar laws and were less apt to take Google and Meta’s threats seriously.

If you’re Google and Meta, this may not seem fair. Links are meant to drive people to websites, right? News sites are getting traffic through those links they otherwise may not have gotten, and the platform loses eyeballs when people click away from it. Meta contends that it doesn’t even post the links in the first place; its users, including the outlets themselves, do that. In the eyes of Google and Meta, they’re doing news sites a favor. And, Meta has said, news content is a very small draw for its users. If the companies don’t really need news links to attract users, why should they be forced to pay for them and be subject to government regulation, something they want to avoid at all costs?”

“In the eyes of the law’s supporters, however, Google and Meta’s business models have taken a lot away from journalism, and this “link tax” is the least they can do to pay some of that back. And, yes, the internet has decimated the journalism industry. One way is digital ad revenues: They’re a fraction of what news outlets commanded for their print and broadcast products, and that already smaller sum is reduced even further because online advertising companies — an industry dominated by Meta and Google — take a cut of it for themselves. One oft-cited statistic has Google and Meta getting 80 percent of online advertising revenue in the country. While Google and Meta have programs that pay news companies, including in Canada, they’re not legally required to do it, they can pick and choose who and what to support (and, by extension, who and what not to support), and they can change the terms whenever they want. Meta, for example, ended an emerging journalists fellowship program in Canada in response to C-18’s passage. The Online News Act is meant to ensure that even the smallest publications get something and that the DNIs have to pay at all. The Canadian government estimates the law will generate about CA$330 million a year for its news outlets.
But that’s all if there are links to Canadian news outlets on those platforms in the first place, which brings us to the current game of chicken between the Canadian government and Big Tech — and the yawning gaps on the news feeds of people like Fenlon and Krichel.”

The Supreme Court decides not to break the internet

“Both Justice Clarence Thomas’s unanimous opinion in Twitter v. Taamneh and the Court’s brief, unanimous, and unsigned opinion in Gonzalez v. Google show admirable restraint. The justices add clarity to a 2016 anti-terrorism law that, if read broadly, could have made tech companies whose products form the backbone of modern-day communications liable for every violent act committed by the terrorist group ISIS.
Instead, the Court’s Twitter and Google decisions largely ensure that the internet will continue to function as normal, provided that websites like Twitter or YouTube do not actively provide assistance to terrorism.

The cases involve similar facts. Google concerns a wave of murders ISIS committed in Paris — one of the victims of those attacks was Nohemi Gonzalez, a 23-year-old American student who died after ISIS assailants opened fire on the café where she and her friends were eating dinner. Twitter, meanwhile, involves an ISIS attack on a nightclub in Istanbul, in which 39 people were killed including Nawras Alassaf, a Jordanian man with American relatives.

At this point, you’re probably wondering what these horrific acts have to do with tech companies like Google or Twitter. The answer arises from the US Justice Against Sponsors of Terrorism Act (JASTA), which permits lawsuits against anyone “who aids and abets, by knowingly providing substantial assistance” to certain acts of “international terrorism.”

The plaintiffs in both cases, relatives of Gonzalez and Alassaf, essentially allege that Twitter, Facebook, and YouTube (which is owned by Google) provided substantial assistance to ISIS by allowing it to use the companies’ social media websites to post videos and other content that promoted ISIS’s ideology and sought to recruit individuals to their cause. In effect, the plaintiffs argued that these tech platforms had an affirmative duty to stop ISIS from using their websites, and that the tech companies could be held liable if ISIS terrorists use a service that is freely available to billions of people across the globe.

It’s a breathtaking legal argument. As Thomas writes in the Twitter opinion, “under plaintiffs’ theory, any U.S. national victimized by an ISIS attack could bring the same claim based on the same services allegedly provided to ISIS.” The three tech companies, in other words, would potentially be liable for any American or relative of an American who is killed by ISIS.

The JASTA statute, moreover, authorizes a successful plaintiff to recover three times the loss inflicted upon them by a terrorist, which in a case similar to Twitter or Google could mean three times the cost of a mass murder. So even a corporate behemoth like Google could potentially be brought to its knees by the amount of money they would have to pay out in future cases if these lawsuits had prevailed.

The Court’s unanimous opinion, however, rejects that outcome. Though the plaintiffs’ theory rests on a plausible reading of the vaguely worded JASTA statute, the Court’s decision establishes that, at the very least, a company has to do more than provide its product to any customer in the world — including customers who may use that product for evil purposes — in order to be held liable for a terrorist act.”

Biden’s ‘Buy American’ Rules Are Getting in the Way of Biden’s Rural Broadband Push

“The Biden administration has framed its new, tighter “Buy American” regulations as a way to bolster domestic manufacturing and benefit parts of the country that have been left behind by technological innovation.
To many of those same communities, the White House has promised better connectivity and higher internet speeds. The bipartisan infrastructure plan signed by President Joe Biden in 2021 dedicated $42 billion to expanding broadband access, with much of the funding aimed at laying fiber optic lines in parts of the country where they don’t exist.

There’s one small problem with all this: Finding enough fiber optic cables that comply with the Buy American rules.”

“Under Biden’s Buy American rules, 55 percent of the component parts of any product used in a federal construction project must be sourced in the United States. That disqualifies any imports of finished cable, but it also wipes out most of the available American-made supply since many of the component parts are sourced overseas.”

“Another problem, according to a Bloomberg report earlier this week, is that building a fiber optic network requires more than just fiber optic cable. You also need switches, terminals, routers, and other pieces of tech that are mostly imported or manufactured with imported components. In both cases, the Buy American requirements mean broadband companies can’t use those parts for projects funded with federal funding from the infrastructure bill.

That means less infrastructure gets built, and lots of perfectly good American-made fiber optic cable doesn’t get purchased, simply because less than 55 percent of its components happened to come from somewhere else.”

Government Spending Billions To Expand Broadband but Can’t Tell Who Needs It

“the government currently has no idea where broadband actually is and is not available.
The government defines broadband as any high-speed internet connection that is always on without needing to dial up.”

“To determine what areas need investment, the government relies on maps from the Federal Communications Commission (FCC). But despite costing $350 million, the FCC’s maps are notoriously unreliable and have been for many years. In 2021, The Washington Post noted the maps are based on census data, so “if even one household in a census block—a statistical area that conveys population data—has broadband available, then the agency considers the entire group served. In rural areas, one block could cover dozens of square miles.” The FCC’s maps also don’t take into account physical impediments, like trees and mountains, which can disrupt wireless signals.

As Karl Bode noted this week at Techdirt, the FCC’s maps were so unreliable that multiple states took it upon themselves to draw up their own. Vermont determined that more than 18 percent of its residents lack broadband access, while the FCC’s newly redrawn maps put Vermont’s shortfall at only 3 percent.

Now, with more than $40 billion in state grants on the line, states are scrambling to challenge the new maps, which cost the FCC nearly $45 million in addition to the $350 million previously spent.”

Elections Are Too Online

“In theory, voting machines are already offline, even air-gapped. In practice, however, “many polling places around the country transmit voting results to their county election offices via modems embedded in or connected to their voting machines,” The New York Times reported in 2018, and that’s a point of internet access. Independent investigators in 2019 said they found “nearly three dozen backend election systems in 10 states connected to the internet,” including systems in swing states Wisconsin, Michigan, and Florida—just a “few” weak points. The nonpartisan National Election Defense Coalition says the “assertion that voting machines or voting systems can’t be hacked by remote attackers because they are ‘not connected to the internet'” is a “myth” and has called for results to be transmitted by offline methods, like USB sticks.

That sort of tool would work because the proposal here isn’t that we return to paper ballots in a wooden box or hand-written voter rolls. Paper and the trail it leaves have an important place in electoral security, but I’m not suggesting a completely nondigital approach. We can still have machines as the main counting mechanism, a useful timesaver in uncontested races. Likewise, election authorities can continue to manage voter databases with computers.

Think 1990, not 1890—there’s no need to go full Luddite. But we should disconnect our voting processes from the internet where it’s feasible. We already know online voting is insecure, and given the detrimental effects even small hacks could have on Americans’ confidence in our election outcomes, we’d be wise to harden electoral targets against digital attacks.”