Tech Brew Ride Home - Tue. 04/28 – Deepfakes As A Killer (Advertising) App

Episode Date: April 28, 2020

Shopify unveils its own shopping app. DJI unveils the Mavic Air 2. The Pixel Buds 2 finally arrive. Tesla pushes out a new auto driving update. What really accounts for Magic Leap’s crash and burn? And why deepfakes are probably going to be in a bunch of commercials soon. Sponsors: Metalab.co Plume.com/techmeme Links: Shopify launches Shop, a new mobile app (TechCrunch) NHS rejects Apple-Google coronavirus app plan (BBC News) DJI’s new Mavic Air 2 has an upgraded camera and much longer flying time (The Verge) GOOGLE PIXEL BUDS REVIEW: SECOND TIME’S THE CHARM (The Verge) Tesla says cars can automatically stop for traffic lights (Associated Press) Magic Leap's $2.6 billion bait and switch (TechCrunch) Magic Leap is a Tragic Heap (The Blog of Palmer Luckey) An ESPN Commercial Hints at Advertising’s Deepfake Future (NYTimes) Learn more about your ad choices. Visit megaphone.fm/adchoices

On April 4th, 2023, around 2 in the morning, a man was found stabbed multiple times on a sidewalk in downtown San Francisco. Hey, who did this to you? What happened next turned the story into a political firestorm. Reports have identified the victim as Bob Lee, the founder of Cash App. From Bloomberg Podcasts, this is Foundering, the Killing of Bob Lee, beginning April 16. Welcome to the TechMeme right home for Tuesday, April 28th, 2020. I'm Brian McCullough today. Shopify unveils its own shopping app. DJI unveils the Mavic Air 2. The PixelBuds 2 finally arrive. Tesla pushes out a new auto-driving update, what really accounts for magic leaps, crash, and burn, and why deep fakes are probably coming to a bunch of commercials soon. Here's what you miss today in the world of tech. Speak of the devil. Shopify has gone in. in an interesting new direction today, launching shop, a single consumer shopping app that allows

users to browse recommended and local products, all from Shopify merchants, and all while being able to buy via Shopify's one-click shop pay checkout process. I saw a whole bunch of Shopify people who I started following recently hinting about this all weekend, and I guess this is what they were talking about, quoting TechCrunch. The app is actually an update and rebrand. of Arrive, an app for tracking packages for Shopify merchants and other retailers, which the company says has been used by 16 million consumers already. Shop includes those same package tracking capabilities, but it also allows consumers to browse a feed of recommended products,

learn more about each brand, and make purchases using the one-click shop pay checkout process. Carl Rivera, the general manager of shop, told me that the app is a response to a broader shift, not just from desktop to mobile commerce, but also from mobile web to native mobile apps. The challenge he suggested is that most of us only download and shop from a handful of native apps, so it can be hard for an independent brand to launch an app of its own. What we want to do with shop is give them a place to call their own, Rivera said, end quote. Which is interesting, for all of the reasons we talked about this past weekend on that bonus episode on the one hand, I can see that it is not realistic for every Shopify merchant to spin up

their own mobile app, and it's even less realistic to expect all of us to download a million different shopping apps just to shop. But at the same time, this does seem to smack more of an Amazon-style single shopping platform, doesn't it? Although Shopify says shop recommends products only based on brands you've already shown interest in or have actually bought from in Shopify's stores. And starting with delivery tracking as a gateway use case, if you will, does seem pretty smart in terms of driving adoption quickly. This contradicts the approach in Germany that I told you about yesterday. Apparently, the UK's National Health Service says it is making a contact tracing app that will be using a centralized approach, thus issuing the Google Apple

initiative because the NHSX, which is the tech wing of the National Health Service, says this will give them more insight and data into COVID-19 spread, quoting the BBC. Like the authorities and many other countries, NHSX has opted to use wireless Bluetooth transmissions to keep track of each qualifying meeting and has said that the alerts will be sent anonymously so that users do not know who triggered them. It is opted for a centralized model to achieve this, meaning that the matching process, which works out which phones to send alerts to, happens on a central computer server. NHSX believes a centralized system will give it more insight into COVID-19 spread and therefore how to evolve the app accordingly. One of the advantages is that it's easier to audit the

system and adapt it more quickly as scientific evidence accumulates. Professor Christoph Fraser, one of the epidemiologists advising NHSX, told the BBC. The principal aim is to give notifications to people who are most at risk of having gotten infected and not to people who are much lower risk. It's probably easier to do that with a centralized system, end quote. So if you're keeping track, keeping score of this at home, Switzerland, Estonia, and Austria, as well as a pan-European group called the DP3T, are pursuing decentralized standards for their tracing apps, while France and England are, at least for now, stubbornly alone. in the centralized camp.

Two and a half years after unveiling the original Mavic Air, today DJI has taken the wraps off the Magic Air 2 drone with an upgraded camera sensor that can take 48 megapixel photos and shoot 4K video at 60 frames per second. And the new gadget claims 34 minutes of flying time and is shipping May 11th, starting at $799. Quoting The Verge, DJI also touts this as its smartest and safest drone yet. It comes with pre-programmed scene detection modes for photos that include snow, trees, grass, blue skies, sunsets, and sunrises. All of DJI's autonomous flying modes are getting performance updates as well. The new sensor inside the

Mavic Air 2 is a half-inch quad-bayer sensor. It's similar to the 48 megapixel sensor that's been available in smartphones for the past year or two. By default, it captures a 12 megapixel image, but it's possible to snap shots using the full resolution. The sensor is behind a fixed 28mm F2.8 lens. For video, the Air 2 can shoot 4K footage at up to 60 frames per second. It also has HDR capabilities for video up to 4K 30 frames per second and HDR panorama photos. The MAVIC Air 2 is also capable of exporting 8K time lapse videos. However, not all modes will support 8K at launch,

and are expected to be available around the end of June. The Mavic Air 2 is also the first drone to come equipped with AirSense, a technology that detects and warns the drone operator of nearby aircraft. It will be available in North America first, while other regions will begin shipping units with AirSense this summer due to supply chain issues caused by the COVID-19 pandemic, end quote. Back when the pixel phones were refreshed, you might remember that Google announced the $179 second

gen pixel bud wireless earbuds. They were to be Google's first true wireless earbuds, eliminating that fabric tether that the first generation had. And the emphasis at the time was all on the hands-free Google Assistant voice control capabilities. However, they weren't immediately made available. Delayed for six months or something, as Google said at the time. Well, it's been six months, it's spring, and finally the new pixel buds are here. And at the verge, Chris Welch, Chris Welch, says the design is much improved, wireless charging is cool, Google Assistant works well, and the sound quality is very good. However, one small caveat, quote, new to this iteration is a soft stabilizer arc. Think of it like a silicone fin on each earbud that's meant to help the pixel

buds stay anchored in your ears. I had no issues with comfort, but we've been testing a few pairs of PixelBuds and Dieter and Becca both said, the arc caused some soreness during extended listening, end quote. I should describe this. It's not a silicon fin that goes around the outside of your ear. It's actually, it tucks inside your kind of ear canal a little bit. I can see how if your ears aren't shaped correctly, that would be a little uncomfortable.

So I don't know if that's a deal breaker or not. I guess it depends on your ear. And I mean, one advantage of the design is the fact that you don't have anything dangling from your head. If you'll recall, that does mean it sort of looks like you've got Frankenstein monster style bolts in your ears. But I guess that's in a weirder than any other design that's out there. And frankly, the coolest application of these things remains the ability to use Google Translate in real time if you're traveling and encountering foreign tongues. Chris's conclusion in his review was this, quote,

At $179, the pixel buds are priced at the same level as Jabra's Elite 75T earbuds, which I'd still personally pick over these for their more dynamic sound and because you don't miss out on features depending on what phone you have. There's unsurprisingly no pixel buds app for iOS. But if Google Assistant has a decent size presence in your life, you might be willing to make do with the downsides of Google's new earbuds, lacking base and the occasional swell in outside noise for the convenience of being able to call upon it anytime, anywhere, end quote. Tesla has rolled out a new update enabling selective vehicles to automatically stop at traffic lights and stop signs somewhat autonomously. As I've always said, full autonomous driving is pretty cool to me in concept and

everything, but in the long interim before we actually get there, I'm more than happy with autonomous half measures like this, quoting the Associated Press. In a note sent to a group of Tesla owners who were picked to test the stoplight and sign recognition feature, the company said it can be used with the traffic awareness cruise control or auto steer systems. The feature will slow the car whenever it detects a traffic light, including those that are green or blinking yellow. It will notify the driver of its intent to slow down and stop, and drivers must push down on the gear selector and press the accelerator pedal to confirm that it's safe to proceed. The company warns in the note obtained by the Associated Press that drivers must pay attention and be ready to take immediate action, quote, including braking because this feature may not stop for all traffic controls, end quote. The note says that over time, as the system learns from the fleet on the roads, it, quote, will control more naturally, end quote.

Tesla didn't respond to multiple requests for additional details, but the website Electrek reported last week that the new feature is being sent to the wider Tesla fleet as part of an over-the-internet software update for thousands of vehicles. The feature won't come until later in other parts of the world, the website said, end quote. I feel like I gave sort of short shrift to the news that Magic Leap was laying off a thousand folks and pivoting to enterprise markets, whatever that might mean. What has been discussed on the show in the past was how big a deal Magic Leap seemed to be at one point, as proven by how many billions of dollars the company raised from the likes of Andreessen Horowitz, Kleiner Perkins, and Google. Sundar Pichai was on Magic Leap's board at one point. Like, these are not folks to be taken in by undercooked technology, you would imagine.

So what went so wrong that the technology that so wowed so many people, so completely actually sunk like a stone to the degree that it now feels like vaporware? Well, John Evans over the weekend had a piece in TechCrunch that sort of feels like a necessary post-mortem to all of this. In short, the early demos, according to Evans, of the Magic Leap technology, might have indeed been full of wow, but the actual product was never able to miniaturize to the point that that tech would wow consumers. Quote, as the information's read Albergotti revealed more than three years ago, The Beast was Magic Leap's original demo box.

It was everything people said it was. It was stunning, dreamlike, breakthrough technology, and it weighed several hundred pounds. The Beast was followed by the cheesehead, which fit on a human head and showed they could miniaturize the light field signal generator they'd invented, but still weighed tens of pounds, obviously far too heavy for any real-world applications. There are actually pictures of both of these units in the linked C-Net story that I'm quoting from here. The Beast and the Cheesehead help explain the multiple rounds of massive venture investment. But then could Magic Leap miniaturize their breakthrough technology further to anything actually

releasable? Clearly they could not. And that's the crux of the matter. The answer to how and why Magic Leap raised $2.6 billion, then laid off half its employees, while hardly releasing anything at all in seven years. To quote Vanity Fair, quoting the information, quoting CEO Roni Abovitz,

the technology behind the beast is not really what we're ultimately going to be shipping Abavits, told the information, adding that prototypes were merely good for showing investors and others what was good about it, what was not, end quote. Intended or not, I assume it wasn't,

Magic Leap became a $2.6 billion bait and switch, the consequences of which are now all to appear. end quote. Check the piece for a more blow-by-blow in terms of details. And also worth reading Oculus founder Palmer Lucky's original scathing takedown of the Magic Leap technology from back in 2018. His piece was called Magic Leap is a tragic heap. There's a link to that in the show notes as well. Finally today, if you've been watching that Michael Jordan documentary on ESPN, the Last Dance, then you've no doubt seen the commercials with Sports Center anchor, Kenny Maine. The spot in question looks like a clip from a 20-year-old sports center itself.

And if you listen, you can actually tell this is not on the up and up. Let's get right to it. Chicago Bulls have won their second three-peat. This is the kind of stuff that ESPN will eventually make a documentary about. They'll call it something like The Last Dance and make it a 10-part series and release it in the year 2020. It's going to be lit. You don't even know what that means yet. While I'm making predictions, jeans will get extra time. There'll be a thing called a butt fumble, and this clip will be used to promote the documentary in a state farm commercial.

It's just a hunch. Yeah, the commercial makes use of deepfake technology to repurpose old footage of Maine. You know how a lot of people are saying this whole coronavirus is just accelerating trends that were already coming to the surface in the before times. Well, things like deep fakes were on the up in the before times. and with the pandemic halting filming, apparently ad agencies are increasingly turning to digital technology like deepfakes to create something new, anything new since they can't actually film something new, quoting the New York Times. We're so restricted in how we can generate content, said Kerry Hill, the production director for the ad agency FCB in North America. Anything that can be computer generated is something we're going to explore, end quote. Hussani Oakley, the chief technology officer of the ad firm Deutsche,

said digitally altered ads should somehow clue viewers into the fact that what they are seeing is not completely real. The technology is here and it's only going to get better and we have to get used to it, he added. We're exploring ways to have fun with it, end quote. Quote, we tried to make the joke clear enough so that we're not tricking anyone, said Kerry Brzeinski Sue, the head of ESPN Creative Works, which created the commercial with ad agency's optimum sports and translation. Ms. Brasinski-Schoo said, Manipulated footage was likely to appear in future ESPN ads as well, and executives at several major advertising agencies said they had discussed

making similar commercials with their clients in recent weeks, end quote. As promised, here's product review number one, and I should note, none of this is paid for or anything. I just randomly bought this stuff recently. I'm just sharing my experiences. I bought a Roomba for the first time since I, owned, I think it was the first generation Rumba, all the way back in 2004, whenever it was. I ended up getting the cheapest Rumba you could get, the 600 series, just to clean up in the

kitchen and the dining room in the evenings. And all I can tell you is it worked great. Like, I feel like the sweeper system in these things has gotten far more comprehensive than I remember. The device is a lot heavier. Maybe that accounts for it. But also the computer and the room mapping tech seems to be way more intelligent too. The Roomba that I remember, if it got stuck or had an error, it had to start its whole room mapping system all over again from scratch. But the couple times I had to pick up the Roomba last night because it had a problem. I simply picked it up, set it down, and it got its bearings in just a few seconds. So the bottom line is, it was great. And it kept my kids occupied for a good hour last night, which again is worth its weight in

gold these days. I can't imagine how entertaining it would be if we still had a dog with this thing. But anyway, was good. Would recommend Rumba. Talk to you tomorrow.