The monopoly-busting case against Google, Amazon, Uber, and Facebook

Antitrust crusaders have built up serious momentum in Washington, but so far, it’s all been theory and talk. Groups like Open Markets have made a strong case that big companies (especially big tech companies) are distorting the market to drive out competitors. We need a new standard for monopolies, they argue, one that focuses less on consumer harm and more on the skewed incentives produced by a company the size of Facebook or Google.

Someday soon, those ideas will be put to the test, probably against one of a handful of companies. For anti-monopolists, it’s a chance to reshape tech into something more democratic and less destructive. It’s just a question of which company makes the best target.

To that end, here’s the case against four of the movement’s biggest targets, and what they might look like if they came out on the losing end. (Note: Apple was too much of a conventional retailer to make the list, but if you’re wondering what an antitrust lawsuit against Cupertino might look like, this is a pretty good place to start.)


Our best model for tech antitrust is the Department of Justice’s anti-bundling case against Microsoft in the ‘90s, which argued that Microsoft was using its control over the PC market to force out competing operating systems and browsers. If you’re looking for a contemporary equivalent, Google is probably the closest fit. On a good day, Google (or Alphabet, if you prefer) is the most valuable company in the world by market cap, with dozens of different products supported by an all-encompassing ad network. Google also has clear and committed enemies, with Microsoft, Oracle, Yelp, and even the Motion Picture Association of America calling for restrictions on the company’s power.

Some of those restrictions are already starting to take shape in Europe, as Google faces a $5 billion fine for alleged anti-competitive Android bundling and a separate $4 billion GDPR case that alleges stingy opt-out provisions. Last week, Sen. Orrin Hatch called on the Federal Trade Commission to investigate anti-competitive effects from Google’s dominance in online ads and search, hinting that similar regulatory pressure may not be far off in the US.

But according to Open Markets’ Matthew Stoller, the best long-term remedy for Google’s dominance has more to do with Google’s acquisitions. “If you’re looking for a silver bullet, probably the best thing to do would be to block Google from being able to buy any companies,” says Stoller. “Suddenly, you have to compete with Google, you can’t just be bought out by Google.”

That might sound tame compared to Europe’s billion-dollar fines, but it cuts to the core of how Google is organized. The company has acquired more than 200 startups since it was founded, including central products like YouTube, Android, and DoubleClick. The company’s modular structure is arguably a direct result of that buying spree, and it’s hard to imagine what Google would look like without it. More recent buys like Nest have fallen under the broader Alphabet umbrella, but the core strategy hasn’t changed. Would Google still be an AI giant if it hadn’t bought DeepMind? Probably, but everyone involved would have had to work a lot harder.

Even better, anti-monopoly activists would have a bunch of different ways to block those acquisitions. The Department of Justice’s antitrust division hasn’t contested Google’s acquisitions so far, but it could always change its approach. The strongest fix would come from Congress, where Sen. Amy Klobuchar (D-MN) has introduced a bill that would place an outright ban on acquisitions by any company with a market cap higher than $100 billion. (As of press time, Google is worth roughly $840 billion.)

Of course, Klobuchar’s bill doesn’t focus on Google or even tech giants, but Stoller says that kind of blockade would have a unique effect on how big companies shape the startup world. “All of these companies, from Amazon to Facebook to Google, they proactively find their competitors and buy them out,” says Stoller. “This would push VCs and entrepreneurs to truly compete with Google. Right now, their strategy isn’t to do that because they want to get acquired.”


Amazon makes life hard for its competitors — and by now, the company is competing against nearly everyone. The most notorious example is the company’s wholesale pillaging of Diapers.com in 2010, which saw Amazon drop diaper prices by as much as 30 percent and matching Diapers.com’s pricing move for move until the smaller outfit agreed to be acquired. More recently, smaller retailers say they’re being targeted and priced out by generics from Amazon Basics, which benefits from Amazon’s wealth of data on who’s buying what. Since Amazon has the money to out-discount any competitors, there’s not much anyone can do about it. With a laser focus on consumer benefit (usually meaning lower prices), the company has become a major player in nearly every market it enters.

Since the modern antitrust standard is mostly focused on consumer harm, Amazon has largely avoided regulatory scrutiny, making it a prime target for the new generation of policy minds that are focused on how big companies can distort markets. Anti-monopoly lawyer Lina Khan laid out the case against the retail giant in a 2017 article called “Amazon’s Antitrust Paradox,” in which she argued that the Amazon store had become a utility infrastructure that the company was subverting for its own benefit. (The argument seems to have found favor with FTC commissioner Rohit Chopra, who hired Khan in July.)

In that view, the problem is that Amazon the store gives too much advantage to Amazon the manufacturer. And thanks to acquisitions such as Whole Foods and the power of Prime, Amazon the store keeps getting bigger.

But Stacy Mitchell, co-director at the Institute for Local Self-Reliance, says that could be solved with a Microsoft-style antitrust suit, carving Amazon up into distinct parts and setting new rules for each part. “Amazon needs to be broken up so that the platform is separated from its retail and manufacturing operations,” says Mitchell. “The platform needs to be treated like a common carrier, so it’s required to serve all comers equally.”

In short, it would be court-mandated net neutrality for The Everything Store. That would take a pretty aggressive Department of Justice to get us there, but Khan’s analysis is gaining favor in surprising corners of Washington.


Uber might not seem as scary as it did during the Kalanick years, but it’s still the largest single crowd-labor platform and a vital piece of transportation infrastructure in 600 cities across the world. Sitting between one-off customers and independently contracted drivers, there are lots of ways for Uber to subtly manipulate the market for its own benefit. The most notorious method was surge pricing, which added a multiplier whenever the supply of nearby drivers was running low. More recently, Uber switched to upfront pricing, but the company still has near total control over how much a given ride costs, and how much of that money makes it back to drivers.

That would be fine for a normal business, but it might be a bigger problem for Uber. The company has long insisted that drivers are independent contractors, not employees. That means Uber can’t be a monopoly in the Standard Oil sense, but it could be a part of a price-fixing conspiracy, in which an entire industry colludes to raise prices at once. That usually looks like a bunch of companies secretly agreeing not to compete with each other, like when UK supermarkets all agreed to boost milk prices or Apple convinced publishers to sell ebooks at a single rate. In both cases, the companies were found to be in violation of the Sherman Act, and the conspiracy was broken up.

Marshall Steinbaum, research director at the Roosevelt Institute, says the “independent contractor” structure makes Uber uniquely vulnerable to a conventional antitrust case. “The nature of the business is fundamentally a conspiracy among hundreds of thousands of independent businesses,” Steinbaum says.

One customer has already tried to cast surge-pricing as a price fix in civil court, suing over the higher prices paid as a result of the conspiracy. The case was ultimately thrown out because of an arbitration clause in Uber’s Terms of Service (although not before Uber got in trouble for spying on the plaintiff). Still, the Justice Department isn’t bound by Terms of Service, and it could bring the same case any time it likes.

If the case were successful, Uber and other crowd-labor platforms would be faced with a tough choice. If it keeps drivers as independent contractors, it’d be forbidden from any kind of price control and forced into a flat Airbnb-style marketplace as it scraps it out with competing networks. It could escape those limits by recognizing drivers as employees, but that would subject the company to a battery of new requirements on minimum wage, benefits, and workers’ compensation, immediately becoming the largest employer in the country. Either way, Uber would face a lot more limits on how it treats drivers and passengers.

“It would force them to take away either the ability to charge whatever they want or the ability to treat drivers as independent contractors,” says Steinbaum. “Lose either one, and you’ve undermined the power of having a centralized transportation monopoly.”


In some ways, Facebook is the most urgent case. It’s inescapable, opaque, and it wields immense power over the fundamental functions of our society. More than any other tech giant, Facebook’s power feels like an immediate threat and the most plausible first target for congressional action. Sen. Mark Warner (D-VA) has already laid out 20 different measures that would rein in Facebook and other tech giants, ranging from GDPR-style data portability requirements to more carveouts of Section 230.

But while Warner’s measures focus on nudging Facebook toward more responsible behavior, a growing number of critics see the problem as Facebook itself. It may be that a social network with more than 2 billion users is simply too big to be managed responsibly, and no amount of moderators or regulators will be able to meaningfully rein the company in. For those critics, social networks are a natural monopoly, and no amount of portability requirements will ever produce a meaningful competitor to Facebook or a meaningful check on its power.

If that’s true, a classical antitrust breakup (as some have suggested) would seem like the only option. The best example is the breakup of AT&T, which saw the telecom giant’s local phone business split into “baby bells,” each bound by serious geographical and regulatory restrictions. It’s the classic example of how to cut a giant company into smaller companies without disrupting service.

Public Knowledge’s Harold Feld has been thinking hard about how that model might apply to Facebook. Feld is best known as a telecom lawyer, but he’s become increasingly interested in how the telecom model fits into the new platform era. The obvious answer is to approach it like Google: blocking future acquisitions and breaking off side products like WhatsApp and Instagram.

But if the problem is the all-consuming size of the network, splitting off networks may lead to what Feld calls the “starfish problem.” “If you tear up a starfish, the pieces regrow and now instead of one starfish you have five starfish,” says Feld. “If you’re going to split up Facebook, what’s to prevent it from becoming three Facebooks, each one dominant in its particular market segment? That’s a hard problem for antitrust.”

Facebook would be less powerful without WhatsApp and Instagram, in Feld’s view, but it wouldn’t be entirely de-fanged. Facebook Messenger could pick up most of the slack from WhatsApp, while Facebook photo-sharing tools might start to resemble the severed Instagram in response. You could prohibit Facebook from making any products involving photo-sharing or mobile messaging, but even that wouldn’t touch the broader problem of how to govern a universal network.

“It’s not that we shouldn’t think about a breakup,” says Feld. “It’s that we should think about a breakup. You have to consider how you’re going to address these problems.”

For Feld, the only complete fix is a specific platform regulation bill akin to the Telecom Act that spells out a new set of requirements for privacy, moderation, and all the other issues that have dogged Facebook in recent years. That’s a lot for Congress to handle, but there may be no other way through. “We’re not going to solve it all at once,” says Feld. “We need a new and comprehensive law that will address these issues because they’ve come to have an enormous and out-sized impact on our lives.”

A mega-merger in the prison phone industry is in the FCC’s hands

Securus has had more than its share of negative headlines. In the past few years, the company, which provides technology services to prisons and jails, has been slammed by inmates’ families who say they’re charged outrageous prices to phone loved ones. The controversy has extended into video call and email services, two other places the company has staked a claim. In October, the company was hit with a $1.7 million fine for allegedly misleading the FCC during a regulatory maneuver. By May, attention shifted to another scandal, as the company took heat for enabling warrantless cellphone tracking around the country.

It’s against that backdrop that Securus is now moving ahead with a merger that could further consolidate a market already criticized as woefully consolidated. The company, which already claims to service more than 1.2 million inmates in North America, has announced its intention to acquire ICSolutions, a smaller competitor in the industry. While exact market figures are difficult to come by, and Securus has pointed to a handful of smaller businesses that offer similar services, inmate advocates argue that the merger will allow two companies to effectively dominate the market. The only thing standing in the way is the FCC.

The merger requires approval from the agency, and the battle has been playing out in dueling regulatory filings. The Prison Policy Initiative, a nonprofit organization that has fought against onerous prison phone prices for years, has argued that scandal-plagued Securus lacks the requisite “character” requirements for approval, and should be blocked on those grounds. Among other issues, legal director Aleks Kajstura says, the company has been accused by advocates and FCC commissioners of changing the name of a fee to skirt a ban by the agency, and by the FCC of lying to the agency itself. “I feel like that’s a pretty low bar to cross that Securus has failed at,” Kajstura says. (A Securus spokesperson said the company “does not believe it committed any wrongdoing” and entered into the FCC agreement to “expedite” the approval without admitting liability.)

The organization has said consolidation could mean more than 70 or 80 percent of calling services will be operated by two companies, Securus and GTL, the latter of which has already grown by acquiring competitors. With that much market share, they argue, the companies will be able to further tighten their hold on the industry as facilities and customers are forced to agree to whatever costs and terms the companies might demand.

Securus has strongly contested the figures, and in filings, it questioned the organizations’ methodologies, which rely on sparse publicly available records. But Securus, the Prison Policy Initiative points out, has not released its own figures. “They didn’t really put out a number of their own, which makes me think however they calculate, it’s not any better,” Kajstura says.

“Securus is continually working to provide the highest-quality services as affordably as possible, connecting individuals and keeping communities safe,” a spokesperson said in a statement. “Our industry is highly competitive, with over 40 operators vying for contracts — which has driven consumer call rates down by 37 percent over the last five years. Securus’ acquisition of another operator would not harm competition, but would instead allow us to reduce costs and pass on price reductions to consumers, while investing in public safety and security.”

Today, companies like Securus often bargain with prisons and jails for exclusive contracts, and they offer the facilities a cut of the revenue in exchange. Families, advocates say, bear the heaviest burden from those contracts, which can include price structures that benefit the facilities, as advocates say some calls can cost nearly $25 for 15 minutes. Under a new regime, the advocates argue, the companies will be able to dictate even more stringent terms that will further squeeze families of the incarcerated who already complain about the costs of keeping in touch with an incarcerated loved one. “Clearly, the market’s at the point where they’re doing pretty damn well what they please,” Kajstura says.

Current FCC chairman Ajit Pai has not left a track record that bodes well for advocates. As a commissioner at the agency, before assuming the top role, Pai voted against a proposal to regulate the cost of inmate phone calls. The measure was approved, but after taking over, he declined to let the agency defend it in court where parts were successfully challenged by the phone companies. The decision was a severe blow to advocates who worked for years on the rules, and there is no sign of progress toward curbing costs under the current Republican-controlled FCC, which could make a decision on whether to approve the merger soon.

Pai’s FCC did impose a $1.7 million fine against the company for allegedly misleading the agency during a transfer of control proceeding. In a statement at the time, Pai described the incident as “a very serious matter” — but he moved to let the transaction continue. Then-commissioner Mignon Clyburn, dissenting on the decision, described Securus as “a company that has shown it is willing to operate on the bleeding edge of legality when it comes to this agency’s rules.” She described the FCC’s action as “less a slap on the wrist” than “a pat on the back.”

But the advocates have some allies. Massachusetts attorney general Maura Healey has also supported blocking the merger. In a filing with the FCC, Healey writes that “Securus is seeking to avoid any Massachusetts regulatory oversight of its rates and practices.” The merger would leave an effective duopoly in the state, she writes. Securus has argued against the ability of telecom authorities in the state to regulate prices. In an industry where competition is one of the few checks on the companies’ considerable leverage, the merger is causing anxiety.

“Following many years of exorbitant and unfair charges imposed on inmate calling services, it is abundantly clear that we need more competition and oversight from the FCC and the states, not less,” she writes.

How big is too big? The past and future of tech monopolies


There has always been a whiff of world domination to the aspirations of big tech companies: to become a platform for everyone on Earth, to be the “everything store,” to organize the world’s information. Yet as Facebook, Amazon, Google, and others reach unprecedented scale, regulators, activists, academics, and politicians are starting to wonder how big is too big. Lately, there are even signs of growing bipartisan recognition that something should be done, whether it’s Sen. Mark Warner’s detailed proposal for regulating social networks or Sen. Orrin Hatch’s request, following President Trump’s angry and confusing tweets, that the Federal Trade Commission investigate potential anti-competitive practices by Google.

Past antitrust cases provide clues about how lawmakers might proceed, but today’s tech giants pose some novel questions. Are laws made to rein in railroads effective regulatory tools for digital networks? What is the harm, exactly, when products are cheaper or even free? How does the ability of gigantic platforms to acquire or crush any rival affect the economy and society? How do you regulate a company that’s becoming something more akin to an infrastructure?

This week, The Verge is looking at different aspects of the monopoly debate, from the last big tech antitrust battles in the ‘90s to overlooked markets like prison phones to how the anti-monopoly cases against today’s giants might be built.

It’s time to break up Facebook

Tim Wu thinks it’s time to break up Facebook.

Best known for coining the phrase “net neutrality” and his book The Master Switch: The Rise and Fall of Information Empires, Wu has a new book coming out in November called The Curse of Bigness: Antitrust in the New Gilded Age. In it, he argues compellingly for a return to aggressive antitrust enforcement in the style of Teddy Roosevelt, saying that Google, Facebook, Amazon, and other huge tech companies are a threat to democracy as they get bigger and bigger.

“We live in America, which has a strong and proud tradition of breaking up companies that are too big for inefficient reasons,” Wu told me on this week’s Vergecast. “We need to reverse this idea that it’s not an American tradition. We’ve broken up dozens of companies.”

And breaking up Facebook isn’t a new idea. Ever since Mark Zuckerberg bought Instagram and WhatsApp, the idea of undoing those deals has been present at the periphery of the conversation about regulating tech companies. Both were serious burgeoning competitors to the social network, and both acquisitions sailed through without serious government oversight, which was a mistake. Instead of facing competition, Facebook was able to swallow its rivals and consolidate the market.

“I think if you took a hard look at the acquisition of WhatsApp and Instagram, the argument that the effects of those acquisitions have been anticompetitive would be easy to prove for a number of reasons,” says Wu. And breaking up the company wouldn’t be hard, he says.

“What would be the harm? You’ll have three competitors. It’s not ‘Oh my god, if you get rid of WhatsApp and Instagram, well then the whole world’s going to fall apart.’ It would be like ‘Okay, now you have some companies actually trying to offer you an alternative to Facebook.’”

Breaking up Facebook (and other huge tech companies like Google and Amazon) could be simple under the current law, suggests Wu. But it could also lead to a major rethinking of how antitrust law should work in a world where the giant platform companies give their products away for free, and the ability for the government to restrict corporate power seems to be diminishing by the day. And it demands that we all think seriously about the conditions that create innovation.

“I think everyone’s steering way away from the monopolies, and I think it’s hurting innovation in the tech sector,” says Wu.


Antitrust law in America seems to be at an inflection point: after a proud history of aggressive enforcement that saw the breakups of everything from Standard Oil to the original AT&T, the past few decades have been incredibly lax as something called the “consumer welfare standard” has swept the courts. Basically, the consumer welfare standard says the government has to show that a merger will result in increasing prices for consumers before it can stop it.

“I think anyone will agree that [the consumer welfare standard] creates a very challenging screen where most antitrust cases die,” says Wu. “And sometimes the prices will go up, but you can’t prove it because it’s hard to prove. That’s what I’m trying to overthrow.”

There are two problems with the consumer welfare standard in 2018: first, after years of dancing around it, giant corporations and their lawyers have learned to make their arguments about price increases ridiculously technical. This leads to comical misdirection. For example, the judge in the AT&T-Time Warner merger case devoted hundreds of pages to the technical discussion of price increases and paid zero attention to the anticompetitive effects of AT&T prioritizing its own video services over others.

Second, it’s all but impossible to show a consumer price increase when major internet services like Google and Facebook are free. Making a case for breaking up these companies will rely on showing a different type of harm than high consumer prices — something like anticompetitive practices, or that innovative businesses get suffocated when they’re absorbed by their gigantic acquirers.

“There are a subset of cases where the primary harm manifests itself as an innovation loss,” says Hal Singer, an economist and antitrust expert. “We need to attack them with a different standard because the probability of prevailing under the consumer welfare standard is zero.”

One or two companies ruling over segments of the market has historically chilled innovation, says Wu.

“I think some people in Silicon Valley are like, ‘Yeah, competition is for losers.’ If you’re competing with other people, you might have to make compromises or [suggest] it’s better just to have one guy, the right guy, making all of the decisions. That’s what AT&T thought. They were like, ‘Listen. We know the phone system. We know what works. This internet stuff is never going to work.’ I think Facebook is in exactly the same position,” Wu said. “They’re trying to set themselves up as regulated monopolist for the foreseeable future.”

And the chilling effect of Facebook and other tech giants buying up every promising startup is noticeable. “​I think if we have a tech economy entirely premised on the idea that monopolists may one day buy the underlying thing, it really limits what can happen,” says Wu.

“Google and Facebook didn’t start that way. There’s a really profound difference in the kind of innovation you see when people are afraid of disturbing the mothership versus what you do when you sense a real opportunity. No one’s willing to fund [profound innovation] because you’re not going to displace Facebook or Google. So we go around the edges somewhere and try and find some cute little thing that doesn’t bother anybody too much and get bought out.”

And so the movement to break away from the consumer welfare standard is growing. Sometimes called the New Brandeis movement, the idea is that the law should prioritize competition. It’s the same sort of standard EU regulators have been using to crack down on big tech companies; these standards were originally based on the American approach under Brandeis and Roosevelt.

“I think we need to simply ask [if] what a large company is doing is part of the competitive process,” says Wu. “Whether, in fact, they’re destroying the other company on the merits or whether they are exceeding the bounds of what’s considered fair competition. You want competition to be something where the better product wins, and the question is: is the defendant winning because they have a better product, or are they winning because they’re using dirty tricks?”

“When you go inside an agency, and you’re facing real cases, this is actually what you’re doing,” says Wu, who spent time working at the Federal Trade Commission. “They don’t mess with the numbers. They look at ‘Okay, Facebook’s killing Snap. Are they doing that in reasonable ways? They’re copying them. Are they better than them, or are they actually doing this in unfair ways?’”

Singer has a different proposal that’s modeled after how Congress decided to regulate cable TV providers like Comcast from discriminating against channels owned by competitors. If a smaller company can show that it’s being meaningfully impaired from competing effectively due to some discrimination against its products, it would have a case.

“Zappos and diapers.com would not have had to sell out to Amazon if they had a venue to defend themselves,” Singer says. “Amazon was able to bring both those companies to their knees because every party knew there was no protection under the antitrust laws.”

It’s the same with Facebook now, he says. “Facebook sits down with someone and says, ‘We could steal the functionality and bring it into the mothership, or you could sell to us at this distressed price.’”

“There’s really nothing to stop Facebook from swallowing all of these verticals.”


The 1914 Clayton Antitrust Act allows mergers to be studied for anticompetitive effects, and Wu thinks the case against Facebook is easy enough that it could be broken up without changing the consumer welfare standard. There’s the simple fact that the number of competitors in social networking went down due to the acquisitions of Instagram and WhatsApp, and there’s also the idea that the number of competitors in what Wu calls the “attention market” decreased as well.

“The easiest way to do it is to start by breaking off WhatsApp and Instagram so those are separate companies,” says Wu. “Hopefully, those companies try to introduce more privacy-sensitive or otherwise better social networking options. Right now, because they’re all owned by the same place, they’re never really allowed to get at the mothership and be a true replacement for Facebook. I think WhatsApp is in an even better position [than Instagram], frankly, to try to go at it. They’ve got this great messaging service. Everyone loves it.”

But wouldn’t reaching in to break up Facebook be difficult for the government to justify? Wu thinks differently. “Unless you believe that we want one ruling master of all social networking and it should be Mark Zuckerberg… then there’s no good reason not to break it up,” he adds. “What’s the argument against it?”

“These are corporations,” says Wu. “They have subunits. Sometimes corporations divide by themselves. It’s not that dramatic, and there’s been this campaign to say, ‘Oh my god, this would be like the most insane thing ever.’”


But won’t getting bigger and bigger lead companies like Facebook and Google to make mistakes, become slower, and create opportunities for new challengers? That has largely been the belief of the tech industry, which has seen the fortunes of companies like AOL, Myspace, and Yahoo dramatically rise and fall. Basically: won’t the market solve for monopoly all by itself?

“It is true that bigness is a curse and leads a company to become doddering and bad,” says Wu. “But the mythology and the problem is assuming that these companies sort of automatically go away. AT&T had a monopoly for 70 years, and by the ‘50s or ‘60s, they were not a great company anymore. They were incredibly hostile to anything new. They thought they knew everything. They thought the internet was a mistake. They didn’t believe in modems. They didn’t believe in answering machines. They were this enormous doddering company, but nothing could get rid of them.”

Breaking up AT&T created massive opportunities for competitors to enter the market, and Wu says the 1990s-era antitrust case against Microsoft was a big factor in creating the modern internet as we know it.

“A whole generation of companies — Google, Facebook, some of these early companies — they don’t owe everything to antitrust, but they owe a sizable debt to the antitrust law,” he says.

Singer agrees. “It’s just hard for me to buy into this claim that [Facebook and Google] are going to be toppled the same way that Myspace was toppled. I feel like these are end-states. I don’t see what dislodges their dominance, at least not in our lifetimes.”

“If you wait long enough, maybe 100 years, they’ll go away. But we could very well have Facebook — an inefficient, ineffective, obsolete company — hanging around for another 20 years,” says Wu. “I’m just not really sure that’s what we need.”

At its most philosophical, antitrust law is the compromise between socialism and capitalism. The idea is that neither the state nor private corporations should amass unchecked power. “I wouldn’t want to live in a socialist country, and I don’t like living in a country where unaccountable capital faces no real check on its power,” says Wu. “The American Revolution was about resistance to centralized power. The Constitution is about resistance. No one entity should have too much power.”

“I’m a believer in the industry of the common man or common woman in charge of their own destiny, a nation of small business, small concerns, people feeling a sense of opportunity, and I think what deadens that is always excessive, concentrated power, whether in government or in private industry.”

“I think we’re in a time where we need to bring back the controls on bigness.”

The best note-taking apps for class, work, and life

You might want a note-taking app as an artist, to take notes during a meeting to remember important feedback, or to jot down insights from a meaningful panel. Maybe you’re a celebrity who needs to publish a public apology to social media, and you need more space than a typical post affords. Often, we don’t give much thought to which app we’re using. It’s either the first app that’s available or the one everyone says is good.

Still, there’s a small quality-of-life improvement to enjoy when you switch from a mediocre, passable note-taking app to the best one you can find. And wouldn’t it be nice to have one app that fulfills your note-taking needs so that all your notes are in one place? That’s why we looked for the simplest, easiest-to-use note-taking app that’s relatively affordable. It had to be on as many platforms as possible so you can sync notes on whatever device you might have or purchase in the future, whether that’s a Mac, PC, Android phone, or iPhone.

The best note-taking app needs intuitive features and a nice-looking design. Bonus points if it has cloud storage, autosave, offline access, and voice-to-text recognition. I need to be able to search through my notes and organize them if they are really going to be of any use to me.

Finally, with the recent data privacy concerns, it needs to be as secure as possible from peeping advertisers or bad actors. That said, only one of the apps featured on this list (Standard Notes) had an encryption feature.

Considering all the other features, like organization, smoothness, and a design that makes you want to keep writing, here’s the best note-taking app.


Google released its note-taking app in 2013, and critics gave it mixed reviews for not having adjustable formatting and being inconvenient for storing long-form notes.

Google Keep isn’t for storing long-form notes; for that, you should look toward other apps on this list. Instead, it’s the single note-taking app that lets you sync notes anywhere, including through a Chrome extension, on a webpage, and on mobile. When you log in through Gmail, everything’s synced right up. Most of the apps on this list have downloads for desktop versions, but Google Keep opens up in a webpage, meaning you can switch devices easily without delay.

You can add notes quickly, and the interface is both unique and intuitive to navigate. It’s easy to move notes around in the order you want and pin the ones you need to keep in view. There’s also a substantially long text preview of each note, so you can more easily figure out what a note is about before you open it. You can also doodle in plenty of colors and brushes, so it’s like having an Adobe Sketch app inside a note-taking app. The font is solid.

There have been a number of improvements since its release. In 2016, Google Keep got pinned notes, so you can establish a kind of hierarchy in your notes. On mobile, you can archive notes by flicking them to the right as well as swivel notes around the page.

Some have criticized Keep for not synching into Google Drive. So if you comb through your Google Drive files, you’ll see PowerPoint presentations, spreadsheets, and documents, but no notes. Google Keep works better for recording your grocery list and doodling. So in some ways, it’s a good thing those mundane notes don’t clog up your Drive.

9

Verge Score


If you have boycotted Google’s suite of apps for one reason or another or you need something that can handle long-form notes, iA Writer, which varies in cost across platforms, is a good alternative. For its one-time purchase, you get a nostalgia-inducing typewriter font, a night mode, and a smooth, easy-to-use interface with a light blue highlight that keeps the app looking distinct. You can also toggle the app icon into dark mode.

One of my favorite things about iA Writer is that you can press a button on the right to exit out of typing mode, which is way easier than most note-taking apps that make you swipe up or double tap.

Text previews also make plenty of sense and give you a good excerpt of what your note is about. The font is well-spaced out, and it’s satisfying to see the type spread out across the page as you jot down notes. iA Writer is on Windows for $19.99, macOS for $29.99, and iOS for $4.99. While it’s free to download on Android, you’ll need to pay $8.99 to unlock unlimited note sharing. It syncs with iCloud, Google Drive, and Dropbox. It loses some marks for its price tag, which is steeper than many on this list, but you get a solidly built app for your money.

8.5

Verge Score

The apps on this list are either built for crafting long-form scripts on mobile or recording idle shower thoughts. Some of the most well-known brands have been relegated to the bottom of the list because they’re not as user-friendly and have fallen behind the times. Two of my favorite word-processing apps are on this list instead of the top picks because they’re simply not compatible with all major operating systems.

Whether it’s for their minimalistic look, security, compatibility with other apps you’re already using, or something more intangible that attracts you to these apps, these are all decent options.

8

Verge Score

Good Stuff

  • Beautiful interface
  • Dark mode
  • Corkboard for organizing notes

Bad Stuff

  • Expensive price tag
  • Only on iOS, macOS, and Windows; an Android app is in development
  • Complex interface

8

Verge Score

Good Stuff

  • Several different themes
  • Expansive keyboard
  • Set word count goals
  • Sleek design

Bad Stuff

  • Expensive subscription model
  • iOS Notes lookalike
  • iOS and macOS only

7.7

Verge Score

Good Stuff

  • Basic, straightforward app
  • Minimalist
  • Allows encryption
  • Passcode or fingerprint lock

Bad Stuff

  • Interface is similar to iOS Notes
  • No way to sort or pin notes

7.5

Verge Score

Good Stuff

  • New notes appear in a different color
  • You can doodle and sketch
  • Good-looking notebook icons inside the app

Bad Stuff

  • Zoho on Windows is only in beta
  • Text preview cuts off awkwardly
  • You have to make an account on desktop

7.5

Verge Score

Bad Stuff

  • Sign up through WordPress
  • Search function isn’t intuitive

7

Verge Score

Good Stuff

  • Easy to share documents
  • Formatting features

Bad Stuff

  • Tough to manage basic note-taking
  • Need to download Google Drive for additional features

6.8

Verge Score

Good Stuff

  • Many formatting features
  • Ideal for longer text
  • Automatic spell check

Bad Stuff

  • Dated look
  • Loads slowly
  • Not designed for short notes on mobile

6.5

Verge Score

Good Stuff

  • Doodling
  • Voice to text
  • Bold, italics, underlining are available

Bad Stuff

  • Need a Microsoft account
  • Loads slowly
  • You have to go through multiple pages until you can create a new note

6.5

Verge Score

Good Stuff

  • Cute mascot
  • Tons of formatting features in free version
  • Doodling and sketching

Bad Stuff

  • Only on iOS and macOS
  • Costs $1.49 per month to sync notes on all devices

6

Verge Score

Good Stuff

  • Simple interface
  • You can set reminders
  • Sharing options

6

Verge Score

Good Stuff

  • Traveling and planning templates
  • Works on iOS, Android, macOS, and Windows
  • Dark mode

Bad Stuff

  • Asks you what industry you work in
  • Time-consuming to use templates
  • Tries to do too many things in one app

5.5

Verge Score

Good Stuff

  • Built-in on iOS, and Android has a third-party clone
  • You can access notes on desktop through iCloud.com
  • Formatting features
  • Search feature works

Bad Stuff

  • iCloud.com loads slowly
  • Text previews are short and vague
  • Missing a lot of features like dark mode and universal syncing

5

Verge Score

Bad Stuff

  • Once you sign up, it spams you with emails
  • Templates blend in with actual notes
  • You have to tap into the app more than three times to start adding or writing a note

Vox Media has affiliate partnerships. These do not influence editorial content, though Vox Media may earn commissions for products purchased via affiliate links. For more information, see our ethics policy.

How AI-generated music is changing the way hits are made

The idea that artificial intelligence can compose music is scary for a lot of people, including me. But music-making AI software has advanced so far in the past few years that it’s no longer a frightening novelty; it’s a viable tool that can and is being used by producers to help in the creative process. This raises the question: could artificial intelligence one day replace musicians? For the second episode of The Future of Music, I went to LA to visit the offices of AI platform Amper Music and the home of Taryn Southern, a pop artist who is working with Amper and other AI platforms to co-produce her debut album I AM AI.

Using AI as a tool to make music or aid musicians has been in practice for quite some time. In the ‘90s, David Bowie helped develop an app called the Verbasizer, which took literary source material and randomly reordered the words to create new combinations that could be used as lyrics. In 2016, researchers at Sony used software called Flow Machines to create a melody in the style of The Beatles. This material was then turned over to human composer Benoît Carré and developed into a fully produced pop song called “Daddy’s Car.” (Flow Machines was also used to help create an entire album’s worth of music under the name SKYGGE, which is Danish for “shadow.”) On a consumer level, the technology is already integrated with popular music-making programs like Logic, a piece of software that is used by musicians around the world, and it can auto-populate unique drum patterns with the help of AI.

Now, there’s an entire industry built around AI services for creating music, including the aforementioned Flow Machines, IBM Watson Beat, Google Magenta’s NSynth Super, Jukedeck, Melodrive, Spotify’s Creator Technology Research Lab, and Amper Music.

Most of these systems work by using deep learning networks, a type of AI that’s reliant on analyzing large amounts of data. Basically, you feed the software tons of source material, from dance hits to disco classics, which it then analyzes to find patterns. It picks up on things like chords, tempo, length, and how notes relate to one another, learning from all the input so it can write its own melodies. There are differences between platforms: some deliver MIDI while others deliver audio. Some learn purely by examining data, while others rely on hard-coded rules based on musical theory to guide their output.

However, they all have one thing in common: on a micro scale, the music is convincing, but the longer you listen, the less sense it makes. None of them are good enough to craft a Grammy Award-winning song on their own… yet.


Michael Hobe, co-founder of Amper Music.
Photo by Christian Mazza / The Verge

Of all the music-making AI platforms I’ve tried out, Amper is hands down the easiest to use. IBM and Google’s projects require some coding knowledge and unpacking of developer language on GitHub. They also give you MIDI output, not audio, so you also have to have a bit more knowledge about music production to shape the output into an actual song.

Amper, on the other hand, has an interface that is ridiculously simple. All you have to do is go to the website and pick a genre of music and a mood. That’s it. You don’t have to know code or composition or even music theory in order to make a song with it. It builds tracks from prerecorded samples and spits out actual audio, not MIDI. From there, you can change the tempo, the key; mute individual instruments, or switch out entire instrument kits to shift the mood of the song its made. This audio can then be exported as a whole or as individual layers of instruments (known as “stems”). Stems can then be further manipulated in DAWs like Ableton or Logic.

I had Amper generate the clip of music below while cruising around LA in the back seat of my friend’s car. Using my phone, I picked rock as the genre, and then, appropriately, “driving” as the mood. It spent about a minute churning away before delivering 30 seconds of audio. The result isn’t radio-ready, but it has chords, a little structure, and it sounds… pleasant. It could easily sit in the back of a YouTube video or an advertisement and no one would guess it was coded, not written.

As someone who makes music, the idea that code can do what I do is freaky. It’s unnerving to think that an algorithm can make a not-terrible song in minutes and that AI is getting in on creative turf we categorize as distinctly human. If AI is currently good enough to make jingly elevator music like the clip above, how long until it can create a number one hit? And if it gets to that point, what does it mean for human musicians?


Taryn Southern showing off IBM Watson Beat.
Photo by Christian Mazza / The Verge

These aren’t questions that Taryn Southern is concerned with. Southern is an online personality who you might know from her YouTube channel or when she was a contestant on American Idol. These days, Southern is interested in emerging tech, which has led to her current project: recording a pop album. Those two things don’t sound like they could be related, but her album has a twist: instead of writing all the songs herself, Southern used artificial intelligence to help generate percussion, melodies, and chords. This makes it one of the first albums of its kind, a collaboration of sorts between AI and human.

Amper was the first AI platform Southern used when beginning her album, and now she also works with IBM Watson Beat and Google Magenta. She views AI as a powerful tool and partner, not a replacement for musicians.

“Using AI, I’m writing my lyrics and my vocal melodies to the actual music and using that as a source of inspiration,” Southern tells me. “I find that really fun, and because I’m able to iterate with the music and give it feedback and parameters and edit as many times as I need, it still feels like it’s mine in a sense.”

To get an idea of how a human can work with AI, look at Southern’s 2017 single, “Break Free.” The SoundCloud audio below is an early export of material from Amper. Compare that to the YouTube video that has the final, released version of the song. Bits of the AI-composed original peek through here and there, but it’s more like seasoning, not the main dish. To transform it into a pop song, Southern made a lot of creative decisions, including switching instruments, changing the key, and, of course, writing and performing the vocals.

Southern originally turned to AI because even though she was a songwriter, she knew “very, very little about music theory.” It was a roadblock that frustrated her to no end. “I’d find a beautiful chord on the piano,” Southern says, “and I’d write an entire song around that, but then I couldn’t get to the next few chords because I just didn’t know how to play what I was hearing in my head. Now I’m able to iterate with the music and give it feedback and parameters and edit as many times as I need. It still feels like it’s mine in a sense.”

This feeling of empowerment is exactly what Amper Music is trying to deliver. “I don’t look at it like artificial intelligence,” Amper co-founder Michael Hobe says. “It’s more of intelligence augmentation. We can facilitate your creative process to cut a lot of the bullshit elements of it. For me, it’s allowing more people to be creative and then allowing the people who already have some of these creative aspects to really further themselves.”

When Hobe says “bullshit elements,” he’s talking about a guitarist not knowing how to orchestrate an instrument they’ve never worked with before, the time spent crafting the velocity of individual drum hits, or simply being faced with writer’s block. Amper isn’t meant to create the next AI superstar; it’s meant to enable musicians. Of course, using AI also has the added benefit of allowing Southern and others with no formal music background to participate in making music. It democratizes the creative playing field so anyone can play what they hear in their head, just like Southern.

I ask Southern what she would say to people who think using AI is cheating. “Great,” she says. “Yes, we are totally cheating. If music is concretely defined as this one process that everyone must adhere to in order to get to some sort of end goal, then, yes, I’m cheating. I am leading the way for all the cheaters.” She laughs, and then pointedly says, “The music creation process can’t be so narrowly defined.”

It’s something to think about. Every time a new technology is introduced and that tectonically shifts the way we create music, there are naysayers. Things like AutoTune, the use of samples and loops, and Digital Audio Workstations were all “disruptors” that we adapted to and are now commonplace tools and methods. AI will probably be next.

The technology’s impact on the music industry as a whole remains to be seen. Will it destroy jobs? How will it affect musical copyright? Will it ever be able to work without a human? But people like Hobe and Southern believe it will ultimately reap positive benefits. Sure, an algorithm making music sounds scary because it mirrors human capabilities that we already find mysterious, but it’s also a compelling tool that can enhance said human capabilities. AI as a collaborator increases access to music-making, it can streamline workflows, and it provides the spark of inspiration needed to craft your next hit single.

“You’re collaborating and working with the AI to achieve your goal,” Hobe says. “It’s not that the AI is just doing its own little thing. It’s all about the process between it and you to achieve that final artistic vision.”

Samsung Galaxy Watch review: iteration over innovation

It was my third time trying and failing to get the new Samsung Galaxy Watch to play Spotify music on my home speakers. Eventually, I gave up and just picked up my phone. That experience largely summed up my time with the Galaxy Watch, Samsung’s latest in its long line of smartwatches.

Don’t get me wrong, the Galaxy Watch is good at displaying and delivering information about my schedule, the weather, and my fitness. But it falls short of addressing my digital and physical well-being. Apps like Bixby aren’t ready for prime time, and stress- and sleep-tracking features alert you to issues, but they have no way of properly addressing them.

In many ways, the Galaxy Watch is an extension of Galaxy smartphones and Samsung’s own ecosystem, rather than a standalone wearable. A new smartwatch that pushes the category forward has been a long-overdue option for Android phone owners. And while the Galaxy Watch is a new wearable, it just isn’t that device.

6.5

Verge Score

Good Stuff

  • Crisp, vibrant OLED display
  • Battery lasts at least two days
  • Activity tracking is kind of useful

Bad Stuff

  • BIxby / S Voice is terrible
  • Not comfortable to wear to sleep

One thing that hasn’t changed about smartwatches is price: they’re still expensive gadgets. The Galaxy Watch I reviewed is a 42mnn Wi-Fi / Bluetooth model that sells for $329, but you can go for a larger 46mm version for $349 that is topped off with a bigger battery and more RAM.

Samsung also sells an LTE version starting at $379 for the 42mm model or $399 for the larger 46mm LTE variant. For comparison’s sake, the entry-level 38mm Apple Watch Series 3 also costs $329, while the larger 42mm goes for $359, and the LTE versions are $399 and $429, respectively.


You can use the Galaxy Watch to pay for things, track up to 40 different physical activities, and track your sleep. You can wear it for days without needing to take it off and charge it. Samsung even claims there are 60,000 watchfaces in the Galaxy Apps store, but the watch already includes a handful of interactive watchfaces, including one that blocks out the hours you have meetings so you can better track your schedule.

Overall, the Galaxy Watch is a great example of show-and-tell by Samsung: it’s a decent-looking wearable with good battery life, a bright OLED display, other spec upgrades, and tons of software features. But at the end of the day, it’s still a smartwatch, which means it has all of the familiar baggage.

There are two unavoidable things about the Galaxy Watch: the beautiful OLED screen and the bezel. On the 42mm variant, Samsung uses a 1.2-inch screen with Gorilla Glass DX+ casing for durability. It’s a small screen, but it’s sharp, vibrant, and adjusts brightness accurately enough so you can look at in direct sunlight or in the dead of night. Around it is a metal bezel that mimics that of diving watches — with a pleasant “click” feedback every time you turn it — and it functions as a navigational tool within the watch’s software.

Together, the display and bezel don’t present the most attractive face, but I don’t have any real complaints about the functionality of the display; if it’s too small for you, bump up to the 46mm variant with an ever-so-slightly larger 1.3-inch panel.

Durability is an essential feature for a wearable, so I’m glad to see that Samsung reinforced the Galaxy Watch with 5ATM waterproofing and MIL-STD-810G certification for durability against drops and shocks. I’ve had the confidence to thrash it about and not worry too much about cosmetic damage.

There’s even an “Eject Water” option in settings that moves water out of the speakers, similar to what you can do on recent Apple Watch models.


If you don’t turn the Galaxy Watch on its side, you might miss two buttons with primary and secondary functions. The bottom one triggers the home screen when you’re looking at another widget or screen or shows you all available apps if you press it from the home screen. On the flip side, the top button serves as a back function, so you don’t have to return to the home screen each time. Alternatively, you can hold it and bring up Samsung Pay. They’re straightforward, but the home screen button can also be mapped to open other apps, therefore doubling its functionality.

The Galaxy Watch’s software and interface are largely unchanged from Samsung’s prior models, but there are a few new features. The major new things involve fitness tracking, including tracking 40 different types of physical activity, sleep tracking, and stress tracking. You can have 20 widgets set for different activities, apps, or controls (like music playback).

The Galaxy Watch is even cross-compatible with iOS (though with limited functionality) and other Android phones. But if you are using a non-Samsung Android phone, be prepared to install at least four Samsung apps on your phone for the watch to function.


Overall, it appears as if Samsung is almost too reliant on apps and plugins that you have to download to make the Galaxy Watch tick. An embarrassing example of this was having to pay for groceries but realizing I had to download a Samsung Pay plugin so my Galaxy Watch could talk to my phone, a Galaxy Note 9. It makes no sense that I still needed to download additional software when I already had a Galaxy device, let alone the latest model.

There’s also Bixby! Well, it’s more like a rebranding of Samsung’s old S Voice voice control system that was on earlier Gear smartwatches. It hasn’t been improved, either: it doesn’t have a natural-sounding voice, it frequently mishears my voice (or not at all), and it generally just isn’t useful. You can remap the home button to bring up Bixby by pressing it twice, or saying “Hi, Bixby,” but the latter method only worked half the times I tried it.


I sleep terribly, so I was curious to see how the watch automatically tracked my slumber and what it could tell me about my sleep patterns. At first, I thought the Galaxy Watch did a great job of accurately determining how long I slept, how long I was motionless, and how much of that time was actually spent in a deep sleep (REM).

But data is just data until it’s interpreted, and based on how efficient my sleep is (I’m assuming it’s just counting REM), the Galaxy Watch would tell me about my sleep “efficiency,” but in no way does it recommend or assist with improving my sleep.

Also, sleeping with a watch on is uncomfortable, no matter how hard you try to ignore the fact that a battery, screen, and sensor sandwich is strapped to your arm. Enabling theater mode means the screen won’t wake me up in the middle of the night, but I can still make out the sensor cluster underneath my wrist, which glows green and can be seen easily in the dark.


The same can be said for stress tracking. If I glance at the watch on my wrist, turn the bezel to the stress widget, and tap “Measure,” I can get a gauged reading of what my stress level is like. It takes this data and compares it to my average stress level for the week, then recommends I partake in a deep breathing exercise. That’s all great, but I don’t actually feel any less stressed out in those moments. I actually felt more stressed out reading the Galaxy Watch’s report about how poorly I slept the night before!

There’s also the issue of misreading my stress levels, like urging me to engage in breathing right before a presentation, when, in reality, I feel relaxed. The takeaway here is that the Galaxy Watch is blending the discrete, personal data that a fitness tracker would prioritize and bringing it to a more casual watch setup. In their current capacity, the sleep- and stress-tracking features aren’t the most useful things on the Galaxy Watch.

Tracking your physical activity is what the Galaxy Watch does best. Running, walking, cycling, hiking, swimming, treadmill, circuit training, deadlifts, push-ups, and leg raises are just a handful of the 40 different trackable exercises. And let me tell you: it’s pretty accurate at tracking them and reminding you to stay active.


The Galaxy Watch alerted me at least once a day when I haven’t been active (usually when I’m sitting at my desk), so I should do five torso twists. Interestingly, it didn’t accept shaking my wrist as exercise; I actually had to get up and do it. Similarly, when I worked out with a press group to test the watch’s circuit training and yoga functions, it kept up with accurate readings of my reps and time spent.

Honestly, the accurate workout training is useful if you’re looking for that sort of data. If the same attention to detail could be directed toward the sleep and stress tracking, then Samsung would have a truly compelling well-being device.

Samsung advertises four days of battery life on the 46mm Galaxy Watch and three days on the smaller 42mm model I’ve been wearing. In my testing, it took two days or so for the Galaxy Watch to learn your usage habits and adjust battery consumption accordingly. For example, I opened but forgot to close the Uber Gear app; 30 minutes later, I got a buzzing notification on my wrist suggesting that I close the app by tapping a “Fix Now” button.

By following all these small prompts that were beckoning me to take care of my battery usage, I was able to get full two days of use (that means sleeping overnight) before I hit the 15 percent mark.


I’ve tried my best to buy into the “watch as a companion device” lifestyle that Samsung has presented with the new Galaxy Watch. But being so reliant on Samsung apps to function — even on a Galaxy smartphone — gives me the impression that not much has changed with Samsung’s smartwatches. They still need to be connected to a phone to be their most useful, and they only reach semifunctional independence as an LTE model.

Being able to track most of your physical activities (even sleep) is a feature that can be appreciated by both users looking to learn more about their bodies and tech enthusiasts who want yet another way to stay connected.

Thankfully, the Galaxy Watch also does a good job of telling me what my day will be like with a morning home screen featuring my calendar alerts, news, weather, and basic fitness data. If I didn’t want to look at the alt-barometer, I could just turn the bezel back until I got to my notifications, tap a text, and finally scroll down to a pre-formatted (or optional voice-dictated) response.

However, at the end of the day, I’m left wanting something more useful, seamless, and powerful than the Galaxy Watch. And as a whole, smartwatches in their current capacity aren’t at that point… yet.

Correction: Post was updated to reflect correct pricing for 46mm BT model.

Vox Media has affiliate partnerships. These do not influence editorial content, though Vox Media may earn commissions for products purchased via affiliate links. For more information, see our ethics policy.

Trump lashes out at Google: all the news about the president’s intensifying feud with Silicon Valley


Trump thinks he’s being censored

President Donald Trump’s latest grievance is with Google, a company he thinks is censoring search results to favor left-wing media outlets. Over the course of the last two days and starting with Trump’s parroting of a Fox News segment about Google search result bias, the president has intensified his criticism of Silicon Valley — not just Google, but Facebook and Twitter as well. The ensuing news cycle has been befuddling to say the least, as reporters, White House staff and members of Congress scramble to make sense of Trump’s increasingly misleading and unfounded claims of persecution. Here’s the latest updates on Trump’s war against the tech platforms.

The Cortana-powered GLAS thermostat isn’t as smart as it looks

You probably haven’t considered how your thermostat looks. Even if you have a modern smart thermostat, it probably looks more like a forgettable appliance than something you would enjoy gazing upon.

The GLAS smart thermostat, built by Johnson Controls, is different. This $319 thermostat actually looks — dare I say — attractive and unlike any other thermostat you’ve probably seen before. Its 5.9-inch, translucent OLED touchscreen virtually disappears when it’s idle, but it provides more information about your home’s climate than a Nest or Ecobee. Also, it looks damn cool.

The GLAS is also one of the first connected appliances to have Microsoft’s Cortana voice assistant built right in. Like the Ecobee 4, which is an Amazon Alexa speaker, the GLAS lets you bark commands at it from across the room to adjust the temperature, control other smart home gadgets, or ask for snippets of information, like weather or news. The GLAS also supports control via Alexa and Google Assistant, but for those, you’ll need to use an Echo device or a Google Home to control the GLAS. Cortana is the only assistant on board the device.

Unfortunately, the GLAS is not as good or complete as the competition. It does basic smart thermostat things, and it has the unique ability to measure the air quality in your home, but it falls short in other areas, like determining when you’re home or away and the app experience.

Basically, you have to really like the looks of the GLAS to justify its price over a Nest or Ecobee.

6.5

Verge Score

Good Stuff

  • Stunning looks
  • Large, easy-to-read display
  • Support for Cortana, Alexa, and Google Assistant
  • Indoor air quality measurement

Bad Stuff

  • High cost
  • No remote sensor support
  • Slow app and frustrating usability issues

Installing the GLAS is similar to installing other smart thermostats, and your mileage may vary depending on what HVAC systems are already in your home. If you have a simple, single-stage heating and cooling system like I do, the installation is just a matter of mounting the GLAS and plugging the wires in. The system will do an auto-scan to see what type of system you have during setup, and you can correct it if there are any errors. (I had to inform the GLAS of two specific wires that it did not auto-detect on my system.)

Mounting the GLAS does require more work than other thermostats. Because the wires are connected at the bottom of the unit and not in the middle like most other thermostats, you’ll have to do more drywall patching and painting to cover up the installation holes from prior thermostats. The GLAS does come with a large backplate that covers up the marks from older thermostats well, and I used it while testing because I hate doing painting and drywall work. But if you really want to take advantage of the seamless look of the GLAS, installing it without the backplate is the way to do it. I wish Johnson Controls had included a built-in level in the mounting hardware, like Nest and Ecobee do, as it would have made installation that much quicker and easier.

The translucent OLED touchscreen is the defining feature of the GLAS, and it’s very well-executed. Though it has a far lower resolution than a smartphone (it’s just 480 x 272 pixels), it’s bright and easy to read. It also, surprisingly, doesn’t get covered in fingerprints when you use it.

The downside of the GLAS display is that it turns off quickly and doesn’t turn back on unless you’re right in front of it. Unlike the Ecobee or the Nest, both of which have always-on display features, you can’t read the GLAS from across the room or even from just a couple feet away. That has proved to be annoying in my home because the location of my thermostat is in an upstairs hallway that’s visible from downstairs. I’m able to see the display on either the Ecobee 4 or the Nest Thermostat from downstairs, but the GLAS requires me to walk up the stairs to see what the current temperature settings are.

The GLAS has sensors on its front to measure temperature, air quality, and whether or not you’re home. But unlike Nest or Ecobee, it doesn’t support remote sensors, so it can only measure the temperature in one location. That makes it difficult to balance the system in homes that have hotter or colder areas in them. It also means the GLAS has only one point of reference for activity in the home, and if you don’t move in front of the thermostat for a period of time, it will aggressively set itself to its efficient away mode. In the week and a half I’ve been testing the GLAS in my home, it has thought nobody was home and turned off the air conditioning on multiple occasions, even though people were home in other parts of the house each time.

Like the Nest and Ecobee, you can tell the GLAS when you wake up, when you go to bed, and specific times for when you might be away at work or school. But it doesn’t do a good job of learning your patterns automatically, and it doesn’t seem to adjust itself based on manual temperature adjustments I made after the initial programming. It just doesn’t seem much smarter than a standard programmable thermostat.

Its ability to measure indoor air quality is unique, however. Its sensors can assess the level of volatile organic compounds (VOCs) and carbon dioxide in the air, in addition to standard humidity. The GLAS then assigns a score to your home’s air quality based on those three factors. If the quality is fair or poor, it will recommend running the fan to circulate air through the home. But it won’t just run the fan automatically when it determines that the air quality is less than good, which seems like something a smart thermostat should do. Instead, I programmed it to run the fan for 15 minutes each hour, which the Nest and Ecobee also allow me to do.

The last unique part of the GLAS is its built-in Cortana integration. It has this because its software is built on the Windows 10 IOT core, so it’s actually Microsoft’s Windows powering this thing. The only visible manifestations are the fact that it has Cortana and uses the Windows 10 virtual keyboard whenever you need to type in things like a Wi-Fi password.

I was able to activate Cortana on the GLAS from across the room by saying “Hey Cortana” and issuing a command, such as “set the air conditioning to 75 degrees.” The speaker on the GLAS won’t win any awards for sound quality, but it’s clear and you can adjust the volume.

The larger problem is that Cortana just isn’t as complete of an in-home virtual assistant as Alexa or Google Assistant, and I didn’t really find myself needing to talk to it much. Cortana has fewer device integrations for smart home gadgets and fewer third-party app or skill integrations. And since I already have Alexa speakers in my home that are easier to access and talk to than my thermostat (the GLAS can be controlled by them or a Google Home), I didn’t even notice when Cortana reset itself a few days into my testing and needed me to log back in.


That wasn’t the only bug I experienced, either. The GLAS disconnected itself from my Wi-Fi network numerous times, requiring me to reset its connection manually from the thermostat. The mobile app (iOS and Android only; there’s no web interface for desktop control) mimics the control scheme of the thermostat and offers the same number of functions, but it’s horribly slow to connect, even when I’m home and on the same Wi-Fi network as the thermostat.


Those kinds of issues make it frustrating to use the GLAS thermostat, especially when both Nest and Ecobee’s systems perform much better. Add those to the fact that the GLAS doesn’t have remote sensor support (Johnson Controls tells me it’s coming, but it won’t say when), doesn’t have as good automatic functions, and costs considerably more than even the most expensive Nest or Ecobee, and it makes the GLAS look rather weak compared to the competition.

Many of its weaknesses were things the first versions of the Nest and Ecobee suffered from, as well. (Nest thermostats just got the ability to use remote sensors this year, for example.) Since this is just the first version of GLAS, hopefully Johnson Controls will be able to iron out the wrinkles and add more features as time goes on.

But where it is today, floating in front of your wall, the GLAS does look pretty damn cool. If only its smarts were as compelling as its looks.

Photography by Dan Seifert / The Verge

Vox Media has affiliate partnerships. These do not influence editorial content, though Vox Media may earn commissions for products purchased via affiliate links. For more information, see our ethics policy.

How should we regulate facial recognition?

Facial recognition is everywhere — airports, police stations, and built into the largest cloud platforms in the world — with few federal rules to govern how it’s used. That’s been true for years, but a string of embarrassing stories in recent months has driven home exactly how dangerous the technology can be in the wrong hands, and it’s led to new calls for regulation. Even Microsoft, one of the largest providers, has called on Congress to place some kind of restriction on how and where the technology can be used.

A truly effective facial recognition law would have to tackle several problems at once. Many facial recognition algorithms still show higher error rates for African-Americans, women, and young people, suggesting the systems might be entrenching societal biases. Beyond bias, the sheer power of facial recognition as a surveillance tool has led some groups to call for a moratorium on police use. And now that the technology is accessible to anyone with a cloud developer account, the private sector privacy issues are even harder to ignore.

That leaves reformers with a difficult question: how can we fix facial recognition? We put the question to five leading figures on both sides of the policy fight.

Alvaro Bedoya, executive director of the Center for Privacy and Technology at Georgetown Law. The Center’s Perpetual Lineup project includes a model bill for regulating facial recognition, focused on restricting police access to driver’s license and mug shot databases.


We regulate commercial privacy on any number of other technologies. Credit cards, we regulate them. E-health records, we regulate them. Your cable viewing habits, protected. Your video rental habits, protected. So this idea that just because it’s a technology, we shouldn’t regulate it doesn’t fit with the entirety of the US commercial privacy landscape. It’s never been an argument in commercial privacy to say to Congress, “Oh, this is a technology. It’s just ones and zeroes. You shouldn’t really regulate it.” And it’s already regulated for 1 out of 8 Americans. In Illinois and Texas, the rule is to get permission. So there’s precedent for it both in how we regulated commercial privacy in the past and how we already regulate face recognition today. And I think the simple rule of getting people’s permission makes a ton of sense.

Beyond that, I’d like to see rules around bias and accuracy testing. You probably want protections for children. It should probably not be used on people who are 18 or younger. You probably want prohibitions on sensitive areas like hospitals or clinics or schools, where even if someone’s consented, you still shouldn’t use it.

Look at surveillance cameras. You do not usually have surveillance cameras in bathrooms. You can have them all throughout the store, you’ll have them at the entrance, the exit, in the aisles, at the checkout. But you do not have them in the bathroom because we all understand that recording people in the bathroom is a bad idea.

Brian Brackeen, CEO of the facial recognition company Kairos, an outspoken advocate for regulation in the industry.


I can’t wait for regulation. We’re big proponents of regulation, going back to the NTIA working group. There is a need for regulation in this space, both for law enforcement and even private companies in certain use cases. We’re not opposed to consent requirements either. Particularly in private sector use, there is a relationship between our customer and the individual, so they should be able to get consent. In our case, we have a lot of banking clients. We’re verifying a transaction using the customer’s face. There’s obviously a relationship there, and there’s a value to the customer in not being the subject of fraud. So for our use case, there’s no problem getting consent. But when you’re Facebook or you’re Amazon and you’re using this for the government or trying to bait-and-switch from photos you already have, that’s a different thing. And that’s why they were so against it.

What we need is the NTIA process with an anvil over our head. We need to say, “We’re doing regulation in the fourth quarter. This is the first quarter, so you have a year to give us best practices.” So let’s get people in a room working through these issues with that requirement over our heads. I do think there is a will for it from both sides, Republicans and Democrats.

Evan Selinger, philosophy professor at the Rochester Institute of Technology. Together with law professor Woodrow Herzog, Selinger has called for a complete ban on the use of facial recognition, in both public and private use cases, out of concern that the technology is being normalized.


The important question to ask is: what does it take to get the public on board with a massive facial recognition infrastructure? The answer is normalization. Get people used to using the technology all the time. Don’t just make them comfortable with facial recognition technology, engineer the desire for it. Create habits that lead people to believe they can’t live without facial recognition tech in their lives. This is what the consumer side of facial recognition technology is doing: making it seem banal and unworthy of concern. By getting people to see facial recognition technology as nothing extraordinary, an argument about value and risk is being made.

Benji Hutchinson, VP of federal operations at NEC America. A leading vendor for federal facial recognition contracts, NEC has resisted calls for federal restrictions on the technology.


We do not believe in a complete moratorium on the technology, and we do not believe that there is a burning need for over-legislation. A lot of the positions that we’re seeing are coming from tech companies that are new to the space. And I would argue that these large companies are a little bit on their heels. They’re trying to do all they can do to make sure that they don’t lose the contracts they have and do the right thing. I don’t fault them altogether, but I don’t necessarily agree with the approach.

People forget the benefits. That gets drowned out. This is a wildly successful technology that’s been used to stop terror attacks. It’s been used to take criminals off the street. It lets us have paperless, frictionless travel when we’re going through airports. It decreases lines and wait times. It makes people’s lives better. And I think those benefits get lost in all the negativity.

Kade Crockford, director of the Technology for Liberty Program at the ACLU of Massachusetts. The ACLU has called for a moratorium on government use of facial recognition.


Our core concern is that policing in the United States today functions without effective oversight or accountability. There’s a real deficit of trust. And in that ecosystem, it’s really hard to see how any legal requirement could be applied in a way that would truly protect people.

If police want to conduct a wiretap, they have to go to a judge and get a wiretap order, which is like a super-warrant. And they take that wiretap order to a phone company. Under federal law, the phone company is not allowed to disclose information absent from that wiretap order. So there’s a kind of built-in protection there. The phone company could be liable if they allow police to wiretap you without an order, and the actual infrastructure that facilitates the surveillance is not in the possession of law enforcement. It belongs to this third party, the phone company. So there’s a built-in check there.

We do not have a parallel track for face surveillance. What law enforcement wants, and what it seems like Amazon is trying to push on police departments nationwide, is the creation of this infrastructure inside the police department. We just don’t have the civil society or governmental infrastructure to ensure that law enforcement would not abuse that.

Brackeen: Police have both the question and the answer. When I’m doing banking and commercial applications, my customers don’t have databases of everybody’s photo, so they can’t just put a camera in the front of a store and know everybody who’s walking by. But if I am the City of Orlando and I have everyone’s driver’s license plus access to the FBI’s database, then I can put a camera on Main Street and know everybody who’s walking by. So the government has an ability that’s beyond other folks.

Then there’s the bias issue. If you have a tool like a stun gun and I told you that stun gun only works on women, then you wouldn’t allow police to use it, for obvious reasons. The tools of policing need to be equitable across all people, and that’s true of AI in general. I want to get to a place where police can use a tool like this and have it be equitable. But I would still have several limitations to that use.

Bedoya: The idea that this technology is being used to catch shoplifters when it’s documented to not work as well on young people and it’s documented to not work as well on African-Americans is just crazy. Every researcher that looks at this finds that there is bias of some sort. It’s crazier in the law enforcement context, but it’s still pretty crazy in the private sector context. So that’s one critical thing I would like to see.

Brackeen: This is not an intractable problem. It’s a data problem, not a problem with the technology. We’re updating our algorithms ourselves, and we can remove bias, at least up to a point. I think the current outrage is, we’re a small company, but we’re doing the work necessary to be better. A company like Amazon has always had the resources to do better. But, in fact, they’re not, and they’re selling to the government.

Hutchinson: If you talk to any big players in the facial recognition space, all of them will say that we have very rigorous testing methodologies. And that includes ethnic individuals in a heterogeneous database. We look for diversity, and we test for it in our algorithms. We spend millions of dollars a year looking for error rates that occur with different types of faces. And those different types of faces can mean a lot of different things. People with different backgrounds, from different regions of the world, different shades, we do all of that testing. We don’t publish a lot of the results, but it is absolutely in our best interest to ensure that it is a low-error algorithm. The fact is, the math is not biased, it’s not racist. If some companies have lower-end algorithms and they haven’t put the R&D into it and they do have higher error rates with certain ethnic groups, that may just be an issue of a poor algorithm.

Crockford: There’s a problem with the algorithms and the training data being biased, but that’s not the only problem with bias in this technology. The other problem is that police are using mug shot files as the comparison database, and mug shots are themselves biased because of the degree to which policing has been enacted in a disproportionate manner throughout history and up to the present.

We see disproportionate arrests of black and brown people in almost every category of minor offense, whether it’s driving with a suspended license or an expired registration, drug possession, petty larceny, trespassing, disorderly conduct. The bias in marijuana arrests is just astonishing. Even today, 90 percent of the people arrested for marijuana offenses in New York are black or brown. And that’s not because white people don’t smoke pot in public in New York City. It’s just because white people are almost never arrested for that crime. So that is a bias, and by using that database and pretending it’s a neutral technology, it codifies that bias.