The chronological proximity of Amazon and Google’s dueling new technology and product launch events on Tuesday and Wednesday of this week was highly unlikely to have been a coincidence. Which company, therefore, reacted to the other? Judging solely from when the events were first announced, which is the only data point I have as an outsider, it looks like Google was the one who initially put the stake in the ground on September 2nd with an X (the service formerly known as Twitter) post, with Amazon subsequently responding (not to mention scheduling its event one day earlier in the calendar) two weeks later, on September 15.
Then again, who can say for sure? Maybe Amazon started working on its event ahead of Google, and simply took longer to finalize the planning. We’ll probably never know for sure. That said, it also seems from the sidelines that Amazon might have also gotten its hands on a leaked Google-event script (to be clear, I’m being completely facetious with what I just said). That’s because, although the product specifics might have differed, the overall theme was the same: both companies are enhancing their existing consumer-residence ecosystems with AI (hoped-for) smarts, something that they’ve both already announced as an intention in the past:
Quoting from one of Google’s multiple event-tied blog posts as a descriptive example of what both companies seemingly aspire to achieve:
The idea of a helpful home is one that truly takes care of the people inside it. While the smart home has shown flashes of that promise over the last decade, the underlying AI wasn’t anywhere as capable as it is today, so the experience felt transactional, not conversational. You could issue simple commands, but the home was never truly conversational and seldom understood your context.
Today, we’re taking a massive step toward making the helpful home a reality with a fundamentally new foundation for Google Home, powered by our most capable AI yet, Gemini. This new era is built on four pillars: a new AI for your home, a redesigned app, new hardware engineered for this moment and a new service to bring it all together.
Of the two companies, Amazon has probably got the most to lose if it fumbles the AI-enhancement service handoff. That’s because, as Ars Technica’s coverage title aptly notes, “Alexa’s survival hinges on you buying more expensive Amazon devices”:
Amazon hasn’t had a problem getting people to buy cheap, Alexa-powered gadgets. However, the Alexa in millions of homes today doesn’t make Amazon money. It’s largely used for simple tasks unrelated to commerce, like setting timers and checking the weather. As a result, Amazon’s Devices business has reportedly been siphoning money, and the clock is ticking for Alexa to prove its worth.
I’m ironically a case study of Amazon’s conundrum. Back in early March, when the Alexa+ early-access program launched, I’d signed up. I finally got my “Your free Early Access to Alexa+ starts now” email on September 24, a week and a day ago, as I’m writing this on October 2. But I haven’t yet upgraded my service, which is admittedly atypical behavior for a tech enthusiast such as myself.
Why? Price isn’t the barrier in my particular case (though it likely would be for others less Amazon-invested than me); mine’s an Amazon Prime-subscribing household, so Alexa+ is bundled versus costing $19.99 per month for non-subscribers. Do the math, though, and why anyone wouldn’t go the bundle-with-Prime route is the question (which, I’d argue, is Amazon’s core motivation); Prime is $14.99 per month or $139/year right now.
So, if it’s not the service price tag, then what alternatively explains my sloth? It’s the devices—more accurately, my dearth of relevant ones—with the exception of the rarely-used Alexa app on my smartphones and tablets (which, ironically, I generally fire up only when I’m activating a new standalone Alexa-cognizant device).
Alexa+ is only supported on newer-generation hardware, whereas more than half (and the dominant share in regular use) of the devices currently activated in my household are first-generation Echoes, early-generation Echo Dots, and a Tap. With the exception of the latter, which I sometimes need to power-cycle before it’ll start streaming Amazon Music-sourced music again, they’re all still working fine, at least for the “transactional” (per Google’s earlier lingo) functions I’ve historically tasked them with.
And therefore, as an example of “chicken and the egg” paralysis, in the absence of their functional failure, I’m not motivated to proactively spend money to replace them in order to gain access to additional Alexa+ services that might not end up rationalizing the upfront investment.
Amazon unsurprisingly announced a bevy of new devices this week, strangely none of which seemingly justified a press release or, come to think of it, even an event video, in stark contrast to Apple’s prerecorded-only approach (blog posts were published a’plenty, however). Many of the new products are out-of-the-box Alexa+ capable and, generally speaking, they’re also more expensive than their generational precursors. First off is the curiously reshaped (compared to its predecessor) Echo Studio, in both graphite (shown) and “glacier” white color schemes:
There’s also a larger version of the now-globular Echo Dot (albeit still smaller than the also-now-globular Echo Studio), called the Echo Dot Max, with the same two color options:
And two also-redesigned-outside smart displays, the Echo Show 11 and latest-generation Echo Show 8, which basically (at least to me) look like varying-sized Echo Dots with LCDs stuck to their fronts. They both again come in both graphite and glacier white options:
and also have optional, added-price, more position-adjustable stands:
This new hardware begs the perhaps-predictable question: Why is my existing hardware not Alexa+ capable? Assuming all the deep learning inference heavy lifting is being done on the Amazon “cloud”, what resource limitations (if any) exist with the “edge” devices already residing in my (at least semi-) smart home?
Part of the answer might be with my assumption in the prior sentence; perhaps Amazon is intending for them to have limited (at least) ongoing standalone functionality if broadband goes down, which would require beefier processing and memory than that included with my archaic hardware. Perhaps, too, even if all the AI processing is done fully server-side, Amazon’s responsiveness expectations aren’t adequately served by my devices’ resources, in this case also including Wi-Fi connectivity. And yes, to at least some degree, it may just be another “obsolescence by design” case study. Sigh. More likely, my initial assumption was over-simplistic and at least a portion of the inference functions suite is running natively on the edge device using locally stored deep learning models, particularly for situations where rapid response time (vs edge-to-cloud-and-back round-trip extended latency) is necessary.
Other stuff announced this week included three new stylus-inclusive, therefore scribble-capable, Kindle Scribe 11” variants, one with a color screen, which this guy, who tends to buy—among other content—comics-themed e-books that are only full-spectrum appreciable on tablet and computer Kindle apps, found intriguing until he saw the $629.99-$679.99 price tag (in fairness, the company also sells stylus-less, but notably less expensive Colorsoft models):
and higher-resolution indoor and outdoor Blink security cameras, along with a panorama-stitching two-camera image combiner called the Blink Arc:
Speaking of security cameras, Ring founder Jamie Siminoff, who had previously left Amazon post-acquisition, has returned and was on hand this week to personally unveil also-resolution-bumped (this time branded as Retinal Vision) indoor- and outdoor-intended hardware, including an updated doorbell camera model:
Equally interesting to me are Ring’s community-themed added and enhanced services: Familiar Faces, Alexa+ Greetings, and (for finding lost dogs) Search Party. And then there’s this notable revision of past stance, passed along as a Wired coverage quote absent personal commentary:
It’s worth noting that Ring has brought back features that allow law enforcement to request footage from you in the event of an incident. Ring customers can choose to share video, and they can stay anonymous if they opt not to send the video. “There is no access that we’re giving police to anything other than the ability to, in a very privacy-centric way, request footage from someone who wants to do this because they want to live in a safe neighborhood,” Siminoff tells WIRED.
Last, but not least (especially in the last case) are several upgraded Fire TVs, still Fire OS-based:
and a new 4K Fire TV Stick, the latter the first out-of-box implementation example of Amazon’s newfound Linux embrace (and Linux-derived Android about-face), Vega OS:
We’d already known for a while that Amazon was shutting down its Appstore, but its Fire OS-to-Vega OS transition is more recent. Notably, there’s no more local app sideloading allowed; all apps come down from the Amazon cloud.
Google’s counterpunch was more muted, albeit notably (and thankfully, from a skip-the-landfill standpoint) more inclusive of upgrades for existing hardware versus the day-prior comparative fixation on migrating folks to new devices, and reflective of a company that’s fundamentally a software supplier (with a software-licensing business model). Again from Wired’s coverage:
This month, Gemini will launch on every Google Assistant smart home device from the last decade, from the original 2016 Google Home speaker to the Nest Cam Indoor 2016. It’s rolling out in Early Access, and you can sign up to take part in the Google Home app.
There’s more:
Google is bringing Gemini Live to select Google Home devices (the Nest Audio, Google Nest Hub Max, and Nest Hub 2nd Gen, plus the new Google Home Speaker). That’s because Gemini Live has a few hardware dependencies, like better microphones and background noise suppression. With Gemini Live, you’ll be able to have a back-and-forth conversation with the chatbot, even have it craft a story to tell kids, with characters and voices.
But note the fine print, which shouldn’t be a surprise to anyone who’s already seen my past coverage: “Support doesn’t include third-party devices like Lenovo’s smart displays, which Google stopped updating in 2023.”
One other announced device, an upgraded smart speaker visually reminiscent of Apple’s HomePod mini, won’t ship until early next year. There was one other announced device, an upgraded smart speaker visually reminiscent of Apple’s HomePod mini, which won’t ship until early next year.
And, as the latest example of Google’s longstanding partnership with Walmart, the latter retailer has also launched a line of onn.-branded, Gemini-supportive security cameras and doorbells:
That’s what I’ve got for you today; we’ll have to see what, if anything else, Apple has for us before the end of the year, and whether it’ll take the form of an event or just a series of press releases. Until then, your fellow readers and I await your thoughts in the comments!
—Brian Dipert is the Editor-in-Chief of the Edge AI and Vision Alliance, and a Senior Analyst at BDTI and Editor-in-Chief of InsideDSP, the company’s online newsletter.
Related Content