IoT news of the week for April 29, 2022

Amazon is using data from Alexa for ads: Researchers have discovered that Amazon’s digital assistant uses data from voice utterances to create demographic profiles that it sells to advertisers. Advertisers apparently don’t have access to your individual voice utterances, but information such as the songs you ask Alexa to play or items you add to a shopping list do get compiled and sent along. This story was honestly shocking to me, even though I knew Amazon was using data gleaned from Alexa to sell me stuff. I suppose I didn’t realize it would sell that information to others. For more, and for instructions on how to opt out, check out this story. (The Verge)

Arm’s new M85 processor is optimized for ML at the edge: I received an update from Arm on its Total Solutions for IoT program announced about six months ago, along with details on a new M-class processor designed for the IoT. Arm will soon license the Arm Cortex-M85 chip that will become the most powerful microprocessor on offer and one designed for machine learning jobs. Many Arm partners have been clustering multiple Cortex-M33 or Cortex-M7 to create GHz-capable chips for IoT use cases, and the M85 option gives chip designers even more power to play with, which means even more capable edge devices. Arm also announced pre-designed chip blocks for IoT gateways and another IP-block for building voice recognition for battery powered devices as part of its Corstone program. (Arm)

The Connectivity Standards Alliance tries to remind people why they should be excited about Matter: The Connectivity Standards Alliance, the group behind the Matter smart home interoperability protocol, published a blog post this week touting the benefits of Matter’s multi-admin feature. This is the feature that purportedly will allow a consumer to buy any Matter-certified device and have that device work with other Matter-certified devices. Most practically it should allow a consumer to connect that Matter-certified device into their smart home ecosystem run by any party, or even multiple parties (think Google and Alexa) and expect the products to seamlessly work. Nothing in this post explains exactly how this will work and there isn’t much new here, so I’m guessing its release is a way to keep consumers excited about the promise of Matter after the standard was delayed last month. (The CSA)

Will Microsoft’s HoloLens hit the battlefield? While everyone is talking up the metaverse, I feel like the world is ignoring the cautionary tale that is Microsoft’s HoloLens. These glasses are designed for augmented reality (AR) as opposed to immersive virtual reality (VR), but they actually deliver value to companies deploying them in factory environments. I’ve spoken with HoloLens customers who use the glasses to train workers and to provide informational displays to plant workers on the floor, for example. But leaked documents show that Microsoft hasn’t been thrilled with HoloLens’ performance, and may be stopping the creation of its third-generation HoloLens devices as employees defect to Meta and other VR shops. Plus, apparently one of its largest clients for HoloLens devices — the U.S. military — isn’t thrilled with the current headsets either, and an accountability watchdog in the Department of Defense says the military may be set to “waste $22 billion” on the devices if the deal continues to progress. (The Register)

Microsoft helps Heinz “ketchup” with its digital transformation: Kraft Heinz has signed a multiyear deal to move its business software to Microsoft’s Azure cloud. What’s interesting about this news is that the food company also plans to use Microsoft’s digital twin technology in 34 of its factories. And it plans to build what the companies call a “supply chain control tower” to monitor all aspects of the Kraft Heinz supply chain and manufacturing process to ensure products get where they are supposed to go. This is the promise of the IoT, but when you’re trying to track the raw materials that go into a product, the manufacturing process, and the shipment of manufactured goods across a huge food conglomerate, the devil is in the details. The idea that Microsoft’s software, sensors, and algorithms will be able to monitor and streamline the production is both awesome and full of hubris. I can’t wait to see how this turns out. (Microsoft)

There were a billion IoT attacks last year? Really? I saw this story and was stunned by the numbers until I started reading further. The headline shouts that there were more than a billion IoT attacks in 2021, which feels shocking. However if you dig into the data from the report by SAM Seamless Network, you discover that “IoT” is a wide category, with most attacks affecting routers, access points, and Wi-Fi extenders (for a total of 80% of attacks analyzed). The remaining devices affected were NAS products (5%), VoIP phones (4%), cameras (3%), and smart home devices (3%). Older network gear and networked cameras are huge culprits when it comes to vulnerabilities because it’s often older gear that was designed when hardcoded passwords and modern security practices weren’t a big focus. The solution is to get old gear out of businesses and homes. Especially small businesses where owners may not be focused on IT. (VentureBeat)

Advances in gaze detection could benefit the smart home: Researchers at Carnegie Mellon have developed a gaze-tracking technology called EyeMU that combines gaze and hand motions into a user interface. The gaze detection is designed for phones, with a user looking at the phone and flicking their wrist or fingers to give a command. This sort of of interface is a good option for noisy environments where voice won’t work or really quiet ones where a spoken command would be disruptive. While the EyeMU software is designed for smartphones, I’d love to see this sort of research head to the smart home, so people can combine gaze detection with gestures or even voice to control what’s happening in their homes. (TechHive)

Kinexon has raised $130M for location-tracking technoloy: Ten-year-old analytics company Kinexon has raised $130 million in funding to help it expand to more industries and more locations. Kinexon makes hardware and software that gathers data from a variety of systems such as GPS, RFID, and barcode readers, and subsequently understands where people or things are in an environment. The Kinexon software then performs analytics on the information to draw insights that are relevant to the business. In manufacturing, it might gather location data for supplies and let a plant manager know when inventory gets low, or it might monitor where equipment is. For sports clients, it might help measure an athlete’s performance or broadcast data about play to fans. Manufacturing customers include Airbus, BMW, and Siemens, which use the company’s hardware and software to understand exactly where items are in their factories. (Kinexon)