IoT news of the week for Nov. 23, 2018

San Diego’s smart city saves millions of dollars a year: Yes, implementing IoT across an entire city is a daunting and expensive endeavor. The gains, however, might be worth it when you consider that the City of San Diego is saving an estimated $3.4 million a year thanks to a wide range of smart sensors and applications. Much of that savings comes from a whopping 4,200 CityIQ sensor nodes and individually metered LED street lights. San Diego now pays for electricity based on the actual usage of each light rather than a flat dusk-to-dawn rate, while the CityIQ systems provide data on open parking spaces, traffic, and pedestrian flow — even on gunshot sounds, which yield location information to first responders. (LightedMag)

6 IoT predictions for 2019: Some of the forecasts in this list are obvious, while some are not so obvious at all. Predicting that more IIoT companies will turn to cloud solutions, for example, is something that a $10 psychic would probably tell you. But the idea that more IoT applications will be available in an app store or marketplace format isn’t something I’d thought of (and neither did my psychic.) (Bsquare)

What Apple got when it bought Silk Labs: We noted Apple’s quiet purchase of Silk Labs on this week’s podcast and figured that its website was worth a look before it goes dark. In 2016, Silk pivoted from making a privacy-centric smart home hub with AI to focusing solely on AI solutions for other companies’ products. Its approach was essentially, you bring the device and we’ll add the smarts. Silk has integrated deep learning neural networks with applications ranging from audio and video to person-detection and face-recognition capabilities that can be applied to nearly any smart device equipped with a microphone or camera. So maybe we’ll see Silk inside an Apple-branded line of HomeKit security devices in the future? (Silk Labs)

BlackBerry drops $1.4 billion on an AI company for IoT: Apple isn’t the only company snapping up AI these days; BlackBerry is in the market, too. It just announced it’s acquiring Cylance in an effort to boost its own Spark IoT communications platform. How so? Cylance’s anti-malware AI actively scans enterprise networks for intrusions to provide robust real-time threat detection, something that’s sorely needed as we bring more IoT devices onto the network. The all-cash deal is expected to close in February. (Business Insider)

Virtual machines at the edge may be a thing: So says Frank Cittadino, chief executive officer of QOS Networks, which specializes in SD-WANs, or Software-Defined Wide-Area Networks. I initially dismissed the article, but this Cittadino quote stuck with me: “Most organizations today don’t have the ability to process big data at the edge because they lack the computing power and resources needed for machine learning.” After reflecting on it, I think he’s right. That doesn’t mean SD-WANs are the holy grail here, since there are other ways to bring more machine learning smarts to the edge. One is Cittadino’s suggestion for virtual machines on edge devices, which would yield faster deployment, more compute power, and quicker access to data, and is something I hadn’t previously considered. (Forbes)

Nodle reaches one million active daily nodes: If you’re not familiar with Nodle, Stacey profiled the company back in 2017. Essentially, Nodle pays developers to include their software in apps, and uses the Bluetooth in smartphones to bridge data between IoT devices and cellular networks. That keeps connectivity costs down for consumers, device makers, and telcos, since the data piggybacks on communications from various other devices. Who does it help? Display advertisers in retail stores and the City of Paris, which gets asset tracking data from public spaces. Nodle now has an alpha version dashboard to show information on its network of nodes, and it’s pretty impressive considering it’s a giant network of devices not made by Nodle. (Nodle)

What is machine learning?: This article is a great high-level look at machine learning (ML) and is presented in easy-to-understand language. Plus it includes a flowchart, and I’ve never met a flowchart I didn’t like. Most people likely understand the concept of ML, but this article gives us all a solid baseline when discussing the subject. And it goes a step further by explaining the differences among the three types of deep learning. I won’t spoil the surprise by telling what they are; take a guess and then see if you were right. (MIT Technology Review)

IoT, the internet of…Thanksgiving: Here in the U.S., Stacey and I are in holiday mode, with this newsletter wrapping up the Thanksgiving week. So it’s perfectly appropriate to highlight this article with some IoT projects that have a Thanksgiving theme. I particularly like the one that uses sound detection to hear when relatives are arguing over the holiday football game, politics, or other polarizing topics that should probably be left alone during this family time of thanks. We appreciate all of the newsletter subscribers and those who listen to the podcast! (IoT for All)

Want to sponsor this newsletter and the podcast? Request our new 2019 media kit today. Contact andrew (at) staceyoniot.com.