Tech Brew Ride Home - Thu. 01/16 - Apple Acquires Xnor.ai

Episode Date: January 16, 2020

Apple has acquired a pretty interesting AI startup, the Chromium-based Edge browser is here, but there is trouble in Mozilla land, Fitbit is first to market with blood oxygen monitoring, the state of the app economy and ‘instant’ weather forecasts from Google. Sponsor: Metalab.co Links: Exclusive: Apple acquires Xnor.ai, edge AI spin-out from Paul Allen’s AI2, for price in $200M range (GeekWire) XNOR.ai frees AI from the prison of the supercomputer (TechCrunch) Microsoft launches Chromium Edge for Windows 7, Windows 8, Windows 10, and macOS (VentureBeat) Mozilla lays off 70 as it waits for new products to generate revenue (TechCrunch) Fitbit quietly enables blood oxygen tracking on its wearables (Engadget) Venture capital slowly seeps outside of Silicon Valley (Axios) App stores saw record 204 billion app downloads in 2013, consumer spend of $120 billion (TechCrunch) Google says new AI models allow for ‘nearly instantaneous’ weather forecasts (The Verge) Subscribe to the ad-free feed here 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 ride home for Thursday, January 16th, 2020. I'm Brian McCullough today. Apple has acquired a pretty interesting AI startup.

The chromium-based edge browser is here, but there's trouble in Mozilla land. Fitbit is the first to market with blood oxygen monitoring, the state of the app economy, and instant weather forecasts from Google. Here's what you missed today in the world of tech. Various places are reporting that Apple has acquired the Seattle-based startup X-NOR AI. which develops low-powered, edge-based AI image recognition tools. The rumored purchase price was $200 million, quoting Venture Beat. The arrangement suggests that XNOR's AI-enabled image recognition tools

could well become standard features in future iPhones and webcams. XNOR AI's acquisition marks a big win for the Allen Institute for Artificial Intelligence, or AI2, created by the late Microsoft co-founder Paul Allen, to boost AI research. It was the second spin-out from AI2's startup incubator following kit.aI, which was acquired by the Chinese search engine powerhouse Baidu in 2017 for an undisclosed sum, end quote. But actually, let's come back to what XNOR AI actually does, because it's interesting, and I can certainly imagine what Apple might be able to do with this. basically XNOR allows developers to use and drop AI-centric code and data libraries into their apps.

You don't have to have any specific proficiency in AI to use their tools. The company motto is AI everywhere for everyone. But the really exciting thing they do involves the edge. Their tools allow for machine learning and image recognition that doesn't need to be supported by the cloud that doesn't need the computational power of a supercomputer behind it. So you essentially can do low-power artificial intelligence. And by low-power, I also mean low-power full-stop, not just low-processing power. X-Nor developed a wireless AI demo camera that can run entirely on solar power. So you've got high-level AI stuff on low-power, cheap, often-portable

devices. Now, first of all, the ability to do a ton of computational stuff on device without punching back to the cloud, that might be attractive to a company like Apple that is interested in preserving user data privacy. Timely acquisition, huh? But also since Exnor allows for powerful vision and speech recognition to run on comparatively underpowered devices, imagine what this could mean for eye devices. We're talking about object recognition, photo sorting, accurate on-device translation, a whole slew of things. Let me quote from this tech crunch piece from 2017, because it paints a pretty good picture. The Farhadi being interviewed here is Ali Farhadi, the senior research manager at AI2, which, as I just said, spun out Exnor. Quote,

Farhadi showed me the fruits of their labor by opening an app on his phone and pointing it out the window. The view of the Fremont cut outside was instantly overlaid with boxes dancing over various objects. boat, car, phone, their labels read. In a way, it was underwhelming. After all, this kind of thing is what we see all the time in blog posts touting the latest in computer vision. But those results are achieved with the benefit of supercomputers and parallelized GPUs. Who knows how long it takes a state-of-the-art algorithm to look at an image and say, there are six boats, two cars, a phone, and a bush, as well as label their boundaries. After all, it not only has to go over the whole scene pixel by pixel, but identify discrete objects within it and their edges,

compare those to known shapes and so on. Even rudimentary object recognition is a surprisingly complex task for computer vision systems. This prototype app, running on an everyday smartphone, was doing it 10 times a second. The arrangement suggests that XNOR's AI-enabled image recognition tools could well become standard features in future iPhones and webcams. Quote, you could leave this running for hours on your phone, for Hattie said, and we haven't even optimized it for battery drain, end quote. It's doing the work of supercomputers, but drawing no more power than a game, and only using a single core of the CPU. A few moments later, they showed me real-time object recognition running on a Raspberry Pi Zero among the simplest and cheapest modern computers available. It was

even up on a hollow lens. They called it Xnore.a.i after the logic gates that power its efficiency, end quote. Now, an interesting side note to this, Exnor had a partnership with Wise, which makes the Wise Home Security cameras. XNOR provided people detection in a manner described similar to what we just heard above for the Wise cameras until mysteriously terminating their contract in November. Maybe now we know why. Downside is, though, apparently owners of Wise cameras right now are reporting that people

detection has already started turning off on their devices. Microsoft yesterday began rolling out its Chromium-based Edge browser for Windows and MacOS, as well as Android and iOS. If you want to find it on the web, go to Microsoft.com forward slash edge. The promised bells and whistles are all there, including tracking protection, extensions, improved performance and themes. But note, quoting Venture Beat, this is Edge 79 stable. Microsoft has yet to lay out which features will ship in Edge 80, Edge 81, and so on.

When asked what to expect from the next few Edge releases, a spokesperson said developers and users should use the beta releases as an indication. If you want to peek at the pipeline, download one of the Edge Insider channels, beta, updated every six weeks, dev, updated weekly, or Canary updated daily. Insider builds can be installed side by side with Edge Stable. Microsoft plans to ship more Consumer Edge features later this year, but this launch, is largely targeted at businesses. That means AAD support, Internet Explorer mode, and Microsoft search in Bing integration. Business and education IT administrators can download offline packages and policies for Chromium Edge. Many businesses already use Chrome and at least one Microsoft

browser, Internet Explorer and or Edge, and Microsoft is hoping these features will get them to switch to using just Chromium Edge, end quote. Speaking of the browser world, things seem to be somewhat more tumultuous. over at Mozilla, which is apparently laid off around 70 staffers. The reason for the layoffs, it seems that the big plans that Mozilla had last year to diversify its revenue stream beyond simply cashing big checks from Google for their share of the search revenue, those plans have not exactly panned out, quoting TechCrunch. In an internal memo, Mozilla chairwoman and interim CEO Mitchell Baker specifically mentions

the slow rollout of the organization's new revenue-generating products as the reason for why it needed to take this action. The overall number may still be higher, though, as Mozilla is still looking into how this decision will affect workers in the UK and France. In 2018, Mozilla Corporation, as opposed to the much smaller Mozilla Foundation, said it had about a thousand employees worldwide. You may recall that we expected to be earning revenue in 2019 and 2020 from new subscription products, as well as higher revenue from sources outside of search. This did not happen, Baker writes in her memo. Our 2019 plan underestimated how long it would take to build and ship new revenue-generating products. Given that, and all we learned in 2019 about the pace of innovation, we decided to take a more conservative approach to projecting our revenue for 2020.

We also agreed to a principle of living within our means of not spending more than we earn for the foreseeable future, end quote. Mozilla had been testing a bunch of subscription-based services like the Firefox private network and a device-level VPN service. that has not launched yet. And one wonders now if it will. Much needed win for Fitbit, which has beaten the Apple Watch to market by being the first wearable device maker to roll out blood oxygen monitoring as a feature to at least some of its wearables, quoting and gadget. Fitbit's devices from recent years have blood oxygen monitoring SPO2 hardware, but they haven't used it until now. Users on Reddit and elsewhere have noticed that their Versa Ionic and Charge 3 devices are suddenly, if sometimes temporarily, providing blood oxygen data to help track health issues like asthma, heart disease, and sleep apnea. It's not clear who's getting first dibs, but it's not dictated by device type or premium memberships.

Fitbit confirmed the rollout to engage it. The devices are using a combination of red and infrared sensors to determine oxygen variation, although the company cautioned that it wasn't a relative feature. You shouldn't see large variations in your sleep if you're healthy, the company said, end quote. To a large degree, this is designed for folks like me who have sleep apnea. And indeed, if you have a Fitbit device, the way you'll know if this new monitoring has been turned on for your device is if you check under the sleep data category of the Fitbit app. We talked about a breakdown of the VC numbers in Europe recently. So let's take a look at how things are going inside the U.S. According to Pitchbook, the West Coast share of total U.S. V.C. deal value dropped to 50.2% in 2019,

down from 62.3% in 2018, making that the lowest share for VC deal volume on the West Coast since 2013. However, 78% of funding still took place in the usual places, California, New York, and Massachusetts, up from 75% going to those three states in 2018. So, if you're trying to square the difference in those two numbers, the disparity there indicates that both deal counts and values crept up or stayed at least steady in every other region of the country, quoting Axios. Philadelphia, Chicago, New York City, and Washington, D.C. saw incremental upticks in the number of deals in 2019. Higher year-over-year growth was seen in Seattle, 10.6 percent, and San Diego, 8.4%.

Still, there's a reason the Bay Area remains dominant. It's high concentration of talent, entrepreneurs, and customers, along with the lion's share of capital, thanks to Blockbuster IPOs, creating new wealth to invest. Investors agree that, even though there are some shoots of progress in other regions like the Midwest, it will take four or five successful exits and more local institutional investors to kickstart the investment engines. Quote, we have to have multiple billion dollar exits and those will beget new companies. An investor said, it takes time, end quote. If you are interested in this sort of thing in investing NVC outside of the valley, look up the weekend bonus episode we recorded with Steve Case a couple months ago.

date of the episode was, I believe, November 2nd. And more number crunching, but this time for the app economy. App Annie reports that consumers downloaded 204 billion apps in 2019, which was up 6% year-on-year. Consumers also spent $120 billion on apps subscriptions and in-app features globally. The average mobile user spends 3.7 hours each day using apps, quoting TechCrunch. According to App Annie, the record growth in mobile downloads in 2019 can be attributed to the growth taking place in emerging markets like India, Brazil, and Indonesia, which have seen downloads soar 190%, 40%, and 70% respectively since 2016. Meanwhile, download growth in the U.S. has slowed to just 5% during that same time, while China saw 80% growth. That doesn't mean users in mature markets aren't downloading apps, only that the growth in year-over-year download numbers is.

is starting to level off. Still, those more mature markets continue to see large numbers of installs with more than 12.3 billion downloads in the U.S. in 2019, 2.5 billion in Japan, and 2 billion in South Korea. The record numbers are notable also, given that App Annie's analysis excludes reinstalls and app updates, end quote. And also notable that $120 billion spent on apps is a more than 200% increase just from 2016. Games still account for the majority of money spent with apps at 72% of that spending pie, while subscriptions account for 28% up from 18% in 2016. App Annie also forecast that the entire mobile app industry will contribute $4.8 trillion to global GDP by 2023. Finally, today, Google says that it is developing new AI models that could allow for what it

is calling nearly instantaneous weather forecasts, something that would be useful in an era of climate change. Apparently, even the best, most sophisticated existing weather forecasting techniques can still take hours to put together, although that extra time does allow for more detailed, granular forecasting. Quoting the verge, however, The work is in the early stages and has yet to be integrated into any commercial systems, but the early results look promising. In the non-peer-reviewed paper, Google's researchers describe how they were able to generate accurate rainfall predictions up to six hours ahead of time at a one-kilometer resolution from just minutes of calculation. The company's researchers

train their AI model on historical radar data collected between 2017 and 2019 in the contiguous U.S. by the National Oceanic and Atmospheric Administration. They say their forecasts were as good as or better than three existing methods making predictions from the same data, though their model was outperformed when attempting to make forecasts more than six hours ahead of time. This seems to be the sweet spot for machine learning in weather forecasts right now, making speedy short-term predictions while leaving longer forecasts to more powerful models. Noah's weather models, for example, can create forecasts up to 10 days in advance. While we've not yet seen the full effects of AI on weather forecasting,

plenty of other companies are also investigating the same area, including IBM and Monstanto. And as Google's researchers point out, such forecasting techniques are only going to become more important in our daily lives as we feel the effects of climate change, end quote. Indeed, this is a direct quote from the researchers, from the paper, which is linked in the show notes, quote, as weather patterns are altered by climate change and as the frequency of extreme, weather events increases, it becomes more important to provide actionable predictions at high spatial and temporal resolutions. Such predictions facilitate effective planning, crisis management, and the reduction of losses to life and property. A DL-based infrastructure can provide predictions within

minutes of receiving new data, allowing them to be fully integrated into a highly responsive prediction service that may better suit the needs of now casting than traditional numerical methods, end quote. That is all for today, and once again, I've got nothing much to add. So here's the usual. You can always follow me on Twitter at Brian MCC, the show subreddit where you can tip me stories is R slash ride home. And the bottom link in the show notes will sign you up to the ad-free feed. Talk to you tomorrow.