IoT news of the week for July 5, 2019

File this under understanding our computer overlords: When I talk to people about AI or computer vision there’s a huge gap in understanding based on our own biology and self-centeredness. We don’t realize that computers don’t “see” like we do, nor do they make intellectual connections like we do. Which isn’t a good or bad thing per se, but it does affect how we design tasks for computers and offers clues as to why sometimes they don’t work as expected. This article does a great job of explaining how computers see and where things break down. (Quanta)

Investing in deep tech: This article by Benjamin Joffee at HAX Accelerator both defines deep tech (basically areas of convergence in two disciplines) and dispels myths about investments in the sector. If you’re an investor or buyer you should obviously check it out, but it also provides a glimpse into the future and tips around how to start evaluating opportunities more broadly. And that’s helpful for everyone. (Medium)

How to opt out of facial recognition at airports: After the Customs and Border Protection agency hack, I am surprised this isn’t getting more attention, but Wired explains how to avoid using a facial scan to board a plane at airports currently using facial recognition to match flyers to their passports. Essentially you just ask them to scan your passport instead. What’s troubling here is the lengths the reporter had to go to in order to discover this information and the high likelihood that if an employee isn’t well-informed they might insist a passenger use the scanner or be denied entry to their flight. There’s a high element of coercion involved in the way facial recognition at airports is being implemented that will lead to privacy violations. Expect a lawsuit. (Wired)

Eve will extend HomeKit Bluetooth product ranges: Apple’s decision to solely support Wi-Fi and Bluetooth radios brings simplicity but also battery life and wireless range concerns. Eve reportedly plans to tackle the latter challenge by launching the Eve Extend this fall. The Bluetooth range extender will accept Wi-Fi signals and route device commands over Bluetooth to up to eight other Eve products. (9to5 Mac) — Kevin Tofel

Amazon updates its presentation software for a better Show: This news is really aimed at developers but it’s worth a mention because it can bring an improved experience to Amazon Echo Show and Spot devices. The Alexa Presentation Layer (APL) 1.1 adds support for animations of current visual objects as well as vector graphics. And now that Amazon has a new 5-inch Echo Show device, APL 1.1 brings auto-scaling to developers so the visual components of their skills will dynamically scale to properly fit a range of screen sizes. (Amazon Alexa Blog) — Kevin Tofel

Is voice computing the hit we think it is? The Information takes on Alexa asking companies such as Uber how many people are using Alexa to order rides. It concludes that voice isn’t the next app store or the money-making future of the web. My hunch is that the article is right about that, but it’s missing the bigger picture. Voice isn’t the equivalent of an app store; it’s the equivalent of the mouse or a graphical user interface. It’s an interaction model for computing. And while people aren’t calling for rides en masse or buying Alexa-enabled bathroom mirrors, that’s probably not the metric we should look at. A better one might be the percentage of voice searches or the number of people who are comfortable talking to their devices. Also, I think this is a shift that will take time and more transparency around data use so that people first understand how their voice data will be used. (The Information)

Or maybe voice isn’t the next user interface; brainwaves are: Well, my privacy worries just went up exponentially. At the Sensors Expo event last week in San Jose, Calif., a startup called Petal showed off people playing video games who were using their brainwaves to control the action. Brainwave-reading headsets aren’t new. At an event back in 2013, I flew a drone using my brainwaves. Rather, the creepy part of this story is when the Google engineer talks about using brainwaves to derive more context from users so that Google Assistant can better predict their needs. (EETimes)

Some real talk about 5G: Also at the Sensor Expo event, a Qualcomm executive spoke about companies and cities building their own 5G networks. That is far more interesting to me than the current consumer-oriented tests that show a few smartphones in a few cities getting gigabit speeds. The Qualcomm executive pointed out that 5G costs mean it won’t be viable anytime soon for widespread use cases in IoT (hello cars), but could be useful for private networks in factories or smart cities. I tend to agree, and have seen both Ericsson and Nokia start offering equipment designed to let a sysadmin of sorts operate a local 5G network. Additionally, new spectrum policy around the CBRS radio bands could let companies buy their own spectrum for their own networks. The potential for a big shift is here, but a lot of pieces have to come together first. (EETimes)

Want to sponsor this newsletter and the IoT Podcast? Space for 2019 is filling up quickly. Contact andrew (at) staceyoniot.com for a media kit.