Alphabet Chief Legal Officer David Drummond is at the center of a media firestorm, following a new Medium post authored by a former Google employee who was engaged in a years-long relationship with Drummond more than 10 years ago.

Though the extramarital affair was originally reported on last fall by the New York Times in the broader context of Google’s permissive workplace culture, the former employee, Jennifer Blakely, says in a new and far more detailed account of their relationship that Drummond was a serial philanderer, leaving his wife for Blakely, then leaving Blakely after a son that he fathered with her for another now-former Google employee.

She adds that Drummond also had “an affair with his ‘personal assistant’ who he moved into one of his new homes.”

We reached out to Google, asking for comment either from the company or Drummond, and have yet to hear back.

It’s an enormously unflattering portrait and it comes at a delicate time for Google, which found itself at the center of the #metoo movement last year, after the Times revealed that another former executive, Andy Rubin, had been awarded a handsome exit package following a sexual misconduct claim that the company reportedly found credible.

In addition to accusing Drummond of neglecting her, Blakely also accuses him of being an absentee parent to a son who she says was very much planned by the couple, writing that “[m]onths or years would go by where he wouldn’t see [their son] or respond to my calls or texts with updates and pictures of him or even ask how he was doing, let alone how he might help out, knowing full well I was alone and in desperate need.”

Elsewhere in her post, Blakely says that following a custody battle over the boy that she won when he was four-and-a-half-years old, “David began providing ample child support. ”

According to several sources familiar with the situation, the situation is more nuanced than Blakely describes and it was not a particularly happy union, though these same individuals acknowledge that much of her account is true. Drummond, who was Google’s general counsel back in 2001 when they met, was married when he began an affair with Blakeley, who’d been hired into Google’s legal department as a senior contracts manager. Though Google merely discouraged managers from having affairs with subordinates at the time (it wasn’t against the company’s code of conduct), when Drummond finally told the company in 2007 that Blakeley was pregnant with his child, she was moved into a role in the company’s sales department, far from the company’s legal department.

Blakely writes she had “zero experience in sales” and that it was a hard transition to make: “I did my best to keep up but I was floundering and became depressed at work.”

She left the company roughly one year later, in 2008. She says she and Drummond and their son were living together by that time, adding that seven months later, he broke up with her via text, saying he was “never coming back” to their shared home.

“And he didn’t,” she writes.

Drummond’s star has continued to rise at Google, where he has now spent 17 years and currently serves as its senior vice president of corporate development, as well as the chief legal officer for Google’s parent company, Alphabet.

Google has made numerous changes in the face of criticisms that male executives at the company have not faced consequences in the past for sexual misconduct at work — criticisms that boiled into anger, then walk-outs by tens of thousands of Google employees.

Among them: Google no longer force employees to settle disputes with the company in private arbitration, including in cases of sexual harassment or assault; employees can now lodge complaints relating to harassment and other misconduct via a dedicated website; an and they can bring colleagues to support them to meetings related to investigations.

Still, Google has never commented specifically on Blakely’s allegations and it will be curious to see if this newest salvo forces its hand.

The affair, and others at the company, don’t seem to have forced dramatic changes to its official code of contact, in any case. According to a section titled: “Friends and Relatives; Co-Worker Relationships,” it cautions employees  that “romantic relationships between co-workers can, depending on the work roles and respective positions of the co-workers involved, create an actual or apparent conflict of interest. If a romantic relationship does create an actual or apparent conflict, it may require changes to work arrangements or even the termination of employment of either or both individuals involved. Consult Google’s Employee Handbook for additional guidance on this issue.”

The top line news is that Apple is making changes to the way that Siri audio review, or ‘grading’ works across all of its devices. First, it is making audio review an explicitly opt-in process in an upcoming software update. This will be applicable for every current and future user of Siri.

Second, only Apple employees, not contractors, will review any of this opt-in audio in an effort to bring any process that uses private data closer to the company’s core processes.

Apple has released a blog post outlining some Siri privacy details that may not have been common knowledge as they were previously described in security white papers.

Apple apologizes for the issue.

“As a result of our review, we realize we haven’t been fully living up to our high ideals, and for that we apologize. As we previously announced, we halted the Siri grading program. We plan to resume later this fall when software updates are released to our users — but only after making the following changes…”

It then outlines three changes being made to the way Siri grading works.

• First, by default, we will no longer retain audio recordings of Siri interactions. We will continue to use computer-generated transcripts to help Siri improve.
• Second, users will be able to opt in to help Siri improve by learning from the audio samples of their requests. We hope that many people will choose to help Siri get better, knowing that Apple respects their data and has strong privacy controls in place. Those who choose to participate will be able to opt out at any time.
• Third, when customers opt in, only Apple employees will be allowed to listen to audio samples of the Siri interactions. Our team will work to delete any recording which is determined to be an inadvertent trigger of Siri.

Apple is not implementing any of these changes, nor is it lifting the suspension on the Siri grading process that it halted until the software update becomes available for its operating systems that will allow users to opt in. Once people update to the new versions of its OS, they will have the chance to say yes to the grading process that uses audio recordings to help verify requests that users make of Siri. This effectively means that every user of Siri will be opted out of this process once the update goes live and is installed.

Apple says that it will continue using anonymized computer generated written transcripts of your request to feed its machine learning engines with data, in a fashion similar to other voice assistants. These transcripts may be subject to Apple employee review.

Amazon and Google had previous revelations that their assistants were being helped along by human review of audio, and they have begun putting opt-ins in place as well.

Apple is making changes to the grading process itself as well, noting that, for example, “the names of the devices and rooms you setup in the Home app will only be accessible by the reviewer if the request being graded involves controlling devices in the home.”

A story in The Guardian in early August outlined how Siri audio samples were sent to contractors Apple had hired to evaluate the quality of responses and transcription that Siri produced for its machine learning engines to work on. The practice is not unprecedented, but it certainly was not made as clear as it should have been in Apple’s privacy policies that humans were involved in the process. There was also the matter that contractors, rather than employees, were being used to evaluate these samples. One contractor described as containing sensitive and private information that, in some cases, may have been able to be tied to a user, even with Apple’s anonymizing processes in place.

In response, Apple halted the grading process worldwide while it reviewed the process. This post and updates to its process are the result of that review.

Apple says that around 0.2% of all Siri requests got this audio treatment in the first place, but given that there are 15B requests per month, the quick maths tell us that though it is statistically insignificant, the raw numbers could be quite high.

The move away from contractors was signaled by Apple releasing employees in Europe, as noted by Alex Hearn earlier on Wednesday.

Apple is also publishing an FAQ on how Siri’s privacy controls fit in with its grading process, you can read that in full here.

The blog post from Apple and the FAQ provide some details to consumers about how Apple handles the grading process, how it is minimizing the data given to data reviewers in the grading process and how Siri privacy is preserved.

Router maker Eero is expanding its focus on subscriptions with a new two-tier system. Eero already had a $10 per month subscription called Eero Plus. It is now called Eero Secure+. The company is adding a cheaper plan with less features for $3 per month.

It seems a bit counterintuitive that Eero is selling software subscriptions. The company is mostly known for its tiny mesh routers that you can put in every room of your house.

Eero originally introduced a subscription back in 2017. It was designed as a sort of Amazon Prime of internet services focused on security. It included family plans to password manager 1Password, VPN service Encrypt.me and antivirus MalwareBytes.

Eero Secure+ is more or less a new name for Eero Plus. It costs $9.99 per month or $99 per year and includes the same services, as well as a few software additions, such as parental controls, filtering of dangerous websites at the network level as well as ad blocking.

Essentially, Eero intercepts DNS queries and blocks the ones to suspicious content — it could be a phishing site, an adult site or an ad network. The company replaces your default DNS with ZScaler’s DNS for that feature.

If you don’t want 1Password, Encrypt.me or MalwareBytes, you can now subscripte to Eero Secure to get those DNS-powered features. It costs $2.99 per month of $29.99 per year.

As a reminder, Amazon acquired Eero in February 2019. Eero promised that its privacy policy wouldn’t change after the acquisition.

Still, paying a subscription for DNS filtering is a bit odd. Some public DNS services, such as Quad9, block access to malicious websites.

And if you’re looking for a fun weekend project, you can buy a cheap Raspberry Pi and play with Pi-Hole, an open source project that basically does everything Eero Secure does. You can also build your own VPN service as you should never trust VPN services. They don’t make you more secure and they can basically see all your network traffic.

Sony announced two new APS-C mirrorless interchangeable lens cameras at a special event in New York today, and the announcements are big news for anyone who’s looking for a small, capable camera that can handle everything from sport shooting to vlogging. The new a6600 flagship takes everything that is great about the a6400 it introduced earlier this year, and adds a big battery boost, in-body stabilization, a headphone jack and real-time eye autofocus for video.

The a6600 otherwise looks nearly identical to the a6400 on paper – it has the same 24.2 megapixel APS-C sensor, ISO sensitivity that expands all the way up to 102400, burst shooting at up to 11 fps, 3k 40p internal HDR recording and the same quality viewfinder. None of that is in any way a criticism, however – these were all excellent specs when they debuted on the a6400 earlier this year, and they’ll serve a6600 owners just as well.

What’s been added should be plenty exciting, however, since it brings the in-body image stabilization (IBIS) that was present on the a6500 but did not make it into the a6400, which was more capable in many other ways. This is a clutch feature for video creators, and that’s a key market for the a6XXX line for Sony, especially with its class-leading reputation for autofocus and high-quality 4K video capture. Ditto the addition of continuous real-time eye AF during movie recording.

Finally, that bigger Z battery (which debuted on the company’s full-frame mirrorless camera line and brought big battery life gains there) will be a boon to video and still shooters alike. The existing battery that powers the a6400, along with the rest of the line, is rated at 360 shots, but that’s very low compared to the a7 III’s 610 shots, and Sony is saying you’ll get over 2x the shooting time with the Z battery in the a6600 compared to its predecessor.

The a6600 retails for $1,400 U.S. for the body alone, and is available in a kit with the 18-135 lens from Sony for $1,800. It’ll be on sale in November.

Sony also unveiled a new entry-level option in its a6XXX line, the a6100. You get a lot of the benefits of the more expensive cameras here, including the same AF system (albeit without real-time Eye AF for movies, which is only on the a6600), and 4K 30p internal recording. You don’t get S-Log or HLG recording options, or HDR, however, and there’s no in-body stabilization. You’re also stuck with the older battery, lower maximum ISO sensitive (51200 in expanded mode) and a lower resolution electronic viewfinder. Still, all told it’s a good bargain – especially since you get Sony’s outstanding real-time object tracking autofocus feature.

The a6100 retails for $750 U.S. for the body alone, and is available in a kit with the 16-50 lens from Sony for $850, or with both the 16-50 and the 55-210 for $1,100. It’ll be on sale in October.

Finally, there are two new lenses which will be hugely beneficial to Sony APS-C camera shooters looking for pro-level options. The E-mount 16-55 F2.8 G gives you a focal range on par with the best glass available on other camera systems, and the E 70-350mm f4.5-6.3 G OSS gives you a really long reach zoom (105-535mm equivalent on a 35mm) with built-in stabilization in a relatively small package. The new 18-55mm lens sells for $1,400 and will be sold in October, and the 70-350mm goes on sale in November for $1,000.

Ahead of the 2020 elections, Facebook today announced it’s tightening requirements for groups buying political ads on the social network. The company last year began requiring advertisers get authorized to run ads about social issues, elections, or politics, which involves advertisers providing identification to confirm who they are and where they’re located — including a U.S. street address, phone number, business email and website matching the email. Starting in mid-September, Facebook says advertisers will now need to submit more information about their organization in order to run political ads.

Failure to submit this information will see their ads paused, the company says.

There are five options for submitting more information, three of which will confirm the advertiser is registered in some way with the U.S. government. This includes submitting a tax-registered organization identification number (EIN); a government website matching an email ending in .gov or .mil; or a Federal Election Commission (FEC) identification number.

By submitting this information, Facebook will label the advertiser a “Confirmed Organization” in its ad archive. The advertiser will also be allowed to use their registered organization name in the ad disclaimers and the “i” icon that appears on the upper-right hand corner of the ad will read “Confirmed Organization.”

For smaller businesses or local politicians who want to run ads about social issues, elections, or politics, they can choose to submit one of two other options instead: an organization name with a verifiable phone number, business email, mail-deliverable address and business website with a domain matching the email; or they can provide no organizational information and use the Page Admin’s legal name on their personal information document.

In either of these two cases, the “i” icon on the ad will read “About this ad” instead of “Confirmed Organization.”

Facebook says this “i” icon is how people will be able to see who is trying to influence them through the ad.

“Now, with one tap, people will not only see information about the ad, but they’ll be able to see the information Facebook confirmed, such as whether an advertiser used an EIN or FEC identification number. This will allow people to confidently gauge the legitimacy of an organization and quickly raise questions or concerns if they find anything out of the ordinary,” the company explained in a blog post announcing the changes.

Despite Facebook’s rule around political ads, the company admitted that there have been several cases where advertisers attempted to put out misleading “Paid for by” disclaimers on their ads.

VICE, for example, demonstrated how easy it was to manipulate the system by placing ads on behalf of VP Mike Pence, DNC Chairman Tom Perez, and the Islamic State, which Facebook approved. Business Insider also ran fake ads pretending to be Cambridge Analytica, which Facebook also approved.

More recently, Facebook banned conservative news outlet The Epoch Times from running ads on its platform because of ad policy violations. The organization had hidden under page names like “Honest Paper” and “Pure American Journalism,” to bypass Facebook’s ad transparency system in order to run some $2 million worth of ads promoting the president and spreading conspiracy theories about his opponents.

In addition to the changes around the advertiser requirements, Facebook is updating its social issues list in the U.S. to include 10 categories, instead of 20 distinct subjects. This doesn’t represent a narrowing focus, but rather makes the categories themselves broader to encompass more topics. For example, the “Civil and Social Rights” category would include sub-topics like freedom of religion, LGBTQ rights, and women’s rights.

The new categories were based on the issue lists in countries who recently held elections, Facebook says.

Facebook also will no longer require some environmentally-focused ads to submit these additional requirements, based on user feedback. This includes ads that “discuss, debate or advocate for environmental issues,” those that encourage recycling, or those highlighting sustainable products.

In the months ahead, Facebook says it will be updating its Ad Library to make it easier to track and compare U.S. presidential candidate spending, and it will expand its policy to prohibit ads that discourage people from voting.

It will turn its attention to Pages, too, by requiring national candidates or elected officials to go through Page Publishing Authorization, to verify their Pages are using real accounts and are based the U.S. Facebook will then begin exposing more information about the Page, including the business or organization behind it.

“We know we can’t tackle these challenges alone. That’s why we’re calling for sensible regulation and working directly with governments, watchdogs and regulators,” the company wrote in the blog post. “While our efforts to protect elections are ongoing and won’t be perfect, they will make it harder for advertisers to obscure who is behind ads and will provide greater transparency for people.”

Hackers will soon be able to stress-test the Facebook Portal at the annual Pwn2Own hacking contest, following the introduction of the social media giant’s debut hardware device last year.

Pwn2Own is one of the largest hacking contests in the world, where security researchers descend to find and demonstrate their exploits for vulnerabilities in a range of consumer electronics and technologies, including appliances and automobiles.

It’s not unusual for companies to allow hackers put their products through their paces. Tesla earlier this year entered its new Model 3 sedan into the contest. A pair of researchers later scooped up $375,000 — and the car they hacked — for finding a severe memory randomization bug in the web browser of the car’s infotainment system.

Hackers able to remotely inject and run code on the Facebook Portal can receive up to $60,000, while a non-invasive physical attack or a privilege escalation bug can net $40,000.

Introducing the Facebook Portal is part of a push by Trend Micro’s Zero Day Initiative, which runs the contest, to expand the range of home automation devices available to researchers in attendance. Pwn2Own said researchers will also get a chance to try to hack an Amazon Echo Show 5, a Google Nest Hub Max, an Amazon Cloud Cam, and a Nest Cam IQ Indoor.

Facebook said it would also allow hackers to find flaws in the Oculus Quest virtual reality kit.

Pwn2Own Tokyo, set to be held on November 6-7, is expected to dish out more than $750,000 in cash and prizes.

Insta360 has quickly established itself as the leader in 360-degree video capture, at least for the consumer market, and its new GO stabilized camera builds on that legacy and extends some of the tech it’s built into the category of more traditional, non-360-degree footage.

The $199.99 GO is truly tiny – it weighs under an ounce, and measures less than two inches tall, by under an inch wide. It’s tiny, and that’s ideal for the use case that Insta360 has in mind for this device – wearing it or mounting it virtually anywhere for capturing quick clips. The GO’s all about quick action grabs, with a 30-second cap on clip recording, which you trigger by pressing the lone control button on the device (a second press stops the clip, unless you let it run the entire 30 seconds).

GO’s design is clearly meant for social sharing, but its secret weapon vs. just using your smartphone or making use of other devices is that it packs Insta360’s FlowState stabilization on board. This is the company’s digital video stabilization feature, which works to great effect in its Insta360 One X 360-degree camera for smoothing out footage so that even in intense action sequences it’s not nausea-inducing.

GO also features a magnetic body, which is designed to work in tandem with a variety of accessories, including backs for securing them unobtrusively to clothing, an underwater housing (the camera itself is IPX4 rated, which means essentially it’s protected from splashes but not meant to be submerged), and mounts for sticking to things like surf boards or vehicles. It can capture clips at resolution of up to 2720×2720, but it crops the image to 1080p (at 25 fps) for export as a result of the stabilization tech.

Shooting modes include a standard 25fps as mentioned, as well as a 30fps time-lapse which can record up to 8 hours (which will output a 9 second video) and a hyper laps mode that can shoot for up to 30 minutes to generate a 5 minute video. It can capture photos, too, exporting square images at 2560 x 2560 resolution, or a number of landscape options reading down from there.

In addition to simplifying capture, the Insta360 GO also hopes to make editing and sharing much easier, too, with its FlashCut auto editing feature. This software tool uses “AI” according to the company, in order to find the best clips (you can even sort by category, ie. ‘food’) you capture throughout the day and then stitch them together in a final edit. You can also fully tweak the edits it provides if you’d rather be a more involved creator.

The biggest limitation, based on just reading the specs and not having had a chance to test this out yet, is that the battery life is rated at around 200 clips per day, based on an average of 20 seconds per clip. But that’s including recharging the camera when not in use using the included Charge Case, which has 2.5 extra charges using its built-in battery. That and the recording limitation could prove challenging to anyone looking to create a lot of content with this camera, but on the other hand, it’s very easy to ensure you have it with you at all times – even when your smartphone isn’t nearby.

At $199.99, the Insta360 GO isn’t exactly cheap – but it does include the Charge Case, a pendant with a magnet you can use to wear it around you neck, a stand, a clip for clothing and a sticky mount for putting it on most smooth surfaces. You can also laser-engrave it if you purchase it directly via Insta360’s website. But after some missed starts for this category like the Google Clips camera, and earlier entrants like the Memoto and Narrative Clip lifelogging cameras, I’ll be curious to see if Insta360’s additional features help this gadget define a category.

Here’s a nice thing from some companies this morning — and it’s got a compelling back story, to boot. Lego this morning announced a new accessibility initiate that will make building instructions for select kits available as braille or text for voice readers, in order to reach builders with blindness and vision impairment.

The service is currently available for free through the Lego Audio Instructions site. It’s still in pilot mode, which mostly means it’s currently limited to four kits, with one each from Classic Lego, Lego City, Lego Friends and Lego Movie 2. The company is currently collecting feedback from the experiences with plans to build out its offerings at some point in the first half of next year.

The idea comes from Matthew Shifrin, a 22-year-old blind Lego enthusiast. He approached the company with the idea after he and a friend worked together to create instructions for kits that he could read.

“I had a friend, Lilya, who would write down all the building steps for me so that I could upload them into a system that allowed me to read the building steps on a Braille reader through my fingers,” he says in a release. “She learned Braille to engage with me and support my LEGO passion, and then spent countless hours translating LEGO instructions into Braille.”

[gallery ids="1874890,1874889,1874888,1874887,1874886,1874885"]

MIT’s Media Lab helped create a software that uses AI to translate visual LXFML data (LEGO Exchange Format Mel Script) instructions into text. The result of those instructions are currently being hosted on the Lego site.

Duolingo is one of the most popular apps on the market for folks who want to learn languages. And now, the service is giving users a way to truly ‘carpe diem’ by offering a course in Latin.

If you’re wondering why Duolingo is offering lessons for a language that hasn’t been commonly used since the 8th Century, that’s totally fair. Here’s the scoop: Latin is the backbone for a variety of romance languages, as well as the language of origin for plenty of words across a number of languages, including English.

Studying Latin gives folks a way to study the infrastructure of a handful of other languages.

Moreover, there are plenty of historians, linguists, and professors who wish to read works like The Aeneid by Virgil or Metamorphoses by Ovid or Meditations by Marcus Aurelius in their original Latin form.

Duolingo uses mini-games to help users learn languages. Mini-courses reward users for finishing a course and take away ‘a life’ (like a videogame) when users make a mistake. The company hopes that the end result of a user finishing a course is that they’re able to read, write, speak and comprehend that language.

Moreover, Duolingo has troves of data around how people actually learn languages, and uses machine learning to tweak and improve its service based on that data.

Duolingo has raised more than $100 million across five rounds of funding since it launched in 2011, according to Crunchbase.

At a small event in Manhattan this week, Fitbit laid out its future for the press. Tellingly, the event was far more focus on the company’s software play, with the big hardware announcement feeling almost rushed at the end.

Along with an increased focus on health care providers and enterprise, much of its revenue strategy will be tied up in Fitbit Premium, a $10 a month subscription service. The offering marks a major shift for a company whose identity has been so closely tied to hardware for its first decade of existence.

The announcement comes a year and a half after the release of Versa. The smartwatch has helped the company begin to right the ship after several quarters’ worth of financial struggle. And while last quarter found Fitbit’s valuation stumbling a bit on the heels of a disappointing performance by the Versa Lite, the company says it continues to be committed to its core hardware offering.

Following the announcement of Fitbit Premium and the Versa 2 smartwatch, we sat down with CEO and co-founder James Park to discuss the company’s path and what the future holds for Fitbit.

The state of Fitbit

Brian Heater: The flow of today’s briefing was different. In previous years, the company’s always led with hardware.

James Park: You noticed that it was pretty conscious, and I think it’s just to reinforce the fact that what we’re working on is not just about hardware anymore. But it’s equally important that the services component is an important part of our strategy, and also an important part of an overall solution, again, people healthier.

Coursera, the online education platform now valued at over $1 billion, has made its name through its popular e-learning programs, working with some 190 universities and big names like Google to bring some 3,600 courses and 14 degrees online, attracting 43 million learners to date. Now, it’s taking a significant step in its development.

First, Coursera has made its first-ever acquisition, of a startup called Rhyme Softworks, which has built a platform that lets developers build hands-on, virtual e-learning projects that can be executed from a user’s internet browser. And along with that, Coursera is launching a new offering called Coursera Labs, which will let educational institutions and industry partners (and eventually enterprises) build their own hands-on learning projects, powered in part by the Rhyme technology.

The deal and the Labs product have been in the works for a while before today, and so some partners have already started to build projects around the offering, like this project from the University of London called Sleuth, to help teach programming skills:

The terms of the deal to acquire Rhyme are not being disclosed — and it’s not clear who its investors were, as Crunchbase lists only a pre-seed round for the startup. But Jeff Maggioncalda, the CEO of Coursera, said that it includes not just Rhyme’s IP and related assets, but its team of six, who will now form the basis of a new R&D office for the company out of Sofia, Bulgaria led by Rhyme’s co-founder and CEO, Namit Yadav. Coursera will also be investing millions in the company’s technology as part of its Coursera Labs initiative.

The idea behind developing Coursera Labs is to bring more flexibility as well as customization into the Coursera platform. Aimed at courses that have strong project and development components to them such as programming, math, or perhaps marketing projects — but not, for example, necessarily a course on 20th century history or Japanese literature — the idea is to bring in more of the tools that have become standard in those disciplines as they are taught elsewhere, and as they are used elsewhere (such as in job applications). Tools that are supported to Labs projects include Jupyter Notebook, RStudio, VS Code, cloud software consoles, and native desktop applications, the company said. Educators building courses can use templates from these or create custom courses for their needs.

The new initiative to offer Labs signals an interesting shift for Coursera in another regard: it will be the first time that the company will have a significant amount of content that will require a person to be “online” in order to use it, a change from previously, when leaners could take work offline to do it. Maggioncalda said however that bandwidth is not an issue — very little of it is required — which will be key considering the growth that Coursera has seen in many emerging markets outside of its home base of the US.

What the change could mean is interesting: Coursera will be able to offer more analytics now to course creators  (and to learners) about how long projects take to complete, what questions give people more pause, and more. It will also give those building online courses a new window into making content specifically for the Coursera platform, rather than work that was originally built for consumption in the class being translated into the Coursera (or indeed more generally e-learning) environment. Important to note, though, that Coursera Labs will be one part of a wider toolkit: there will still be a lot of Coursera courses that do not use it, and do not really need to.

Maggioncalda hinted that there will be another chapter coming in how Labs and Coursera will be developing later in the year that could see the initiative scaling up even more. That could in part be down to extending Labs to Coursera Enterprise users, which is a newer area for the company that provides training for knowledge workers and now has 2,000 customers, Maggioncalda said.

The first Versa was a revelation for Fitbit. After falling flat with the Ionic, the company’s second true smartwatch finally delivered on the promise of high-profile acquisitions, including Pebble and Vector, with a smartwatch the could truly compete with Apple’s. In fact, the product helped jump-start Fitbit’s sputtering stock and appeared to be the first step on its road to recovery.

Announced earlier this year, the Versa Lite was widely regarded as a misfire — something CEO James Park admitted in a recent earnings call. Although it was right to have price point at front of mind, it wasn’t enough to justify the loss of several key features. And honestly, the original Versa’s $200 MSRP seems to be right in the sweet spot for those looking for a lower-cost Apple Watch competitor (that also happens to work with Android).

A year and a half after the first Versa, Fitbit finally returns with the follow-up, the apply named Versa 2. The new version of the device isn’t a radical departure, but looks to be a pretty solid upgrade. I don’t anticipate too many Versa 1 owners making the upgrade here, but there’s a lot to like in the new version.

Confirming earlier leaks, Alexa is a key part of the update. Without its own voice assistant, Fitbit turned to the most logical place, partnering with Amazon. Seems like a mutually beneficial team up for both parties — Amazon finally breaks into the wearables space in a meaningful way and Fitbit gets access to an already immensely popular smart assistant.

Interestingly, the Versa 2 marks the first time Fitbit has built a microphone directly into a device. The hardware was added specifically for the purposes of bringing Alexa to the product. The assistant is accessed with the press of a button for a wide variety of features, from weather to smart home control. Answers arrive on-display in text form.

“We actually had to work closely with Amazon because Alexa wouldn’t work out of the box on these devices,” CEO James Park tells TechCrunch. “One, we’re still trying to maintain a low power profile. Two, we’re trying to make our devices still affordable. So we didn’t integrate a speaker. So we had to work with Amazon closely to make Alexa work without voice output. So that actually took a lot of collaborations between the two companies.”

Among the more practical additions is improved battery life. Fitbit says the Versa 2 can get “5+ days” on a charge. The watch should still be able to get several days, even with the new always-on watch face display enabled.

Sleep features get a bit of an overhaul, as well. Sleep Score is pretty much what it sounds like, attaching a number to the quality of one’s sleep. That factors in things like heart rate, restlessness, time awake and sleep stages. The company is promising a higher-quality breakdown of that information via its new Premium subscription offering. Smart Wake is a welcome addition, as well. The feature will be rolling out to all Fitbit smartwatches, waking up users within a 30-minute window, based on sleep cycle.

Design-wise, well, this thing looks even more like the Apple Watch than its predecessor. Here’s what Park told me on that front:

With phones, it’s like every phone starts to look the same. But for us, we try to blend around design and the square design into what we call the squircle design that tries to capture both one that looks more like a traditional watch piece but still has a squareish form factor to display information. So we think we’ve struck the right balance. And I think whether it looks like an Apple Watch or not is kind of irrelevant. We’re trying to look at the customer experience and try to see what’s best for the user.

What a lot of it comes down to is kind of getting out of its own way. Gone are the days when the devices were flashy status symbols. These sorts of wearables have become far more utilitarian, and, as such, are designed to occupy a similar space as a good, old-fashioned dumb watch.

All models of the watch also feature Fitbit Pay built-in, which includes access to a number of different global transit systems, including New York, Vancouver, London and Taiwan.

The Versa is priced at $200, the same as its predecessor. There’s a premium model at $230, which includes a special strap and a 90-day pass to Fitbit Premium.

Services are the future of Fitbit. That was the overarching message from an event this week in Manhattan. The small gathering of media outlets found the company spending most of its stage time on software and services, leaving a small window at the end to discuss the launch of a pair of hardware products, the Versa 2 smartwatch and Aira Air scale.

After a decade of leading with hardware, Fitbit has no doubt seen the writing on the wall in the wearables category. It’s true that devices continue to see growth, but with Xiaomi and other Chinese manufacturers devouring the low end of the market and Apple utterly dominating smartwatches, Fitbit is looking at other ways to leverage its presence in fitness.

Fitbit Premium will be a cornerstone of that play. At $10 a month or $80 annually, the company is eyeing a revenue streaming beyond device sales and enterprise/healthcare partnerships. It’s a play that echoes similar moves by companies like Apple, which has seen services and content become an increasingly important part of its revenue model as sales on devices like the iPhone continue to stagnate.

What does $10 a month get you? Custom insights. That’s the promise here. Bespoke health and sleep information that goes deeper than what users of Fitbit’s free app get. The core of the offering is nine guided health and fitness programs, with names like Intro to Healthy Habits, Get More Zzz’s, Habits for Restful Sleep, Get Active, Beginner Running, Run Training, Understand Calories, Kick Your Sugar Habit and Kick Your Salt Habit.

Health has always been core to what Fitbit does, and everyone’s moving to a subscription model anyway, so the offering makes sense on the face of it. But in a world where we’ve already got thousands of monthly subscription services vying for a piece of our electronic deposits, the company’s got a tough road ahead of it convincing consumers to shell out an additional $10 a month.

“The quick pitch is that it’s a one-stop shop that gives you all of what you really need to eat better, become more active, sleep better,” CEO and co-founder James Park told TechCrunch. “And it’s a service that’s tightly integrated with your Fitbit device. And so with a user base of over 27 million active users, definitely for our user base, it’s a really compelling solution.”

The offering promises a better breakdown of Fitbit’s new Sleep Score feature, thousands of video and audio workouts and various gamified offerings designed to better help the wearer move. Perhaps more compelling are a handful of content partners, including Headspace, Daily Burn and Yoga Studio by Gaiam that supplement Fitbit’s own premium content. Of course, Fitbit Premium isn’t designed to replace any of these services outright and will instead include a selection of content from each.

As ever, I’ll reserve full judgement until we get hands-on time with the service, but my concern is the same as with many fitness content services. Much of the data collected requires user input. I suspect the reminder of a monthly fee will go a ways toward convincing users to keep up, but as with so many of these, accuracy comes down to one key fact: You can’t lie to a fitness tracker, but you can lie to an app.

“A lot of the value that we’re trying to provide as well is that the programs provide and adapt dynamically without you having to really do anything,” says Park. “And we’re trying to provide more nudges and notifications that remind you, throughout the course of the day or through your program, what things you should be doing.”

Getting people to pay for a premium version of content they’ve been offered for free is always a heavy lift — and I say that as someone who works in publishing. I haven’t seen a magic bullet in the Fitbit Premium pitch so far, but I’d be more than happy to be convinced otherwise.

The service has already been piloted in Australia and New Zealand. It will start rolling out to U.S. users this month and will be available in 17 English-speaking countries this fall, with more languages arriving next year.

The service will be exclusively targeted at Fitbit users at launch, but Park says the plan is to move toward a more platform-agnostic model going forward, a move that could foretell a further move away from a hardware-revenue model.

“Long-term, we see Fitbit Premium as something that can be used with other devices as well, whether it’s an Apple Watch, Garmin, etc.” says Park. “But initially we’re focused on our existing users, because we have a lot of them.”