Last year, we reported that Google’s Fuchsia team had renewed its efforts to support smart speakers. Long story short, the team had experimented with a single speaker, ditched that effort, then “restored” it later on. More importantly, the Fuchsia team was found to be working on multiple speakers, the most notable of which was an as-yet-unreleased speaker equipped with UWB.
[…]In a newly posted code change, the Fuchsia team formally marked all of its speaker hardware as “unsupported” and altogether removed the related code. Among the hardware now unsupported by Fuchsia, you’ll find the underlying SoCs for the Nest Mini, Nest Audio, Nest Wifi point, a potentially upcoming Nest speaker, and some Android Things-based smart speakers.
The Fuchsia team hasn’t shared a reason why its smart speaker efforts were discontinued. One issue that potentially played a role is that the Amlogic A113L chip used in “Clover” – an unknown device that we suspect may be the Pixel Tablet dock – does not meet Fuchsia’s strict CPU requirements. Amlogic’s engineers attempted to work around this issue, seemingly to no avail.
It also doesn’t help Google fired about 20% of the 400 people working on Fuchsia. Since its discovery about six years ago, Fuchsia has been on an upward trajectory, but the massive layoffs and now the end of the smart speakers project, one has to wonder what the future of Fuchsia is going to be. Everything seemed to point at Fuchsia one day taking hold in Android and Chrome OS, but that seems farther away now than ever.
I honestly thought Fuchsia would be a bigger deal for google. They clearly have the resources to follow through assuming executives were actually willing to throw their weight behind the project. This suggests it wasn’t a big goal or at least not a unanimous one. I don’t know what this means for other Fuchsia projects, but it’s not a good showing that the fuchsia smart speaker will get heaped into the projects that google abandoned.
If I had to guess, there is likely at least some element of political infighting involved. The Android team has a pretty huge influence at Google due to being one their main success stories outside of ads, and I can very much see people being unhappy with an upstart potentially encroaching on their turf.
Kind of sounds like Nokia, where they had multiple development teams inside, all trying to get their OS to be the main one… R.I.P. maemo/MeeGo…
anevilyak,
That sounds possible, I’ve heard of things like this happening at microsoft. One division might consider other divisions their enemy because the resources between them can be a zero sum game. This is likely the case between android and fuchsia. For fuchsia to grow, realistically android would get displaced.
Ohhh, that’s the power of the Linux Kernel, since a lot of people already use it and know how to work with it, changing the kernel to Zircon and run Linux/Android from there will move everyone away from it’s comfort zone. Is Starnix the way to go? Will it finally move Android from Linux kernel to Zircon kernel? Intriguing !!!! https://fuchsia.dev/fuchsia-src/concepts/components/v2/starnix
A large company (any large company) is a many-headed creature, sometimes it moves together in one direction, sometimes the various heads compete with another, and sometimes they “war” against another.
The comments on those “strict CPU requirements” seem to be only about the CPUs not having cryptography extensions, and Google not wanting to make an exception. Which highly likely piled onto their decision to cut the Fuchsia team, along with the brief AI assistant apocalypse earlier this year.
dark2,
I’m not clear on the details, but to the extent that Fuchsia requires cpu extensions that render Fuchsia non-viable, then it begs the question why Fuchsia has requirements leading to it;s own existential demise? I was perplexed about this when reading the article and I wonder what the story is behind it.
Obviously android runs on commodity CPUs without such requirements, was the requirement used for misdirection to kill the project? I don’t know if there’s a conspiracy here or not, I just find it weird.
No, just smart speakers. Google, like Amazon, probably decided to pull the plug on most of the AI assistant product line, but was too far along in the development process to outright cancel their own smart speaker. No need for further work or weird exceptions when you don’t expect the product line to still exist in a few years, and definitely no need for Fuchsia’s potentially endless updates in that case.
dark2,
It just doesn’t seem plausible to me that they have this problem only in relation to smart speakers, Surely if Fuchsia is working fine in other contexts, they should be able to get it working in smart speakers too. It just has the feeling of being a manufactured problem.
What problem? Seems like a classic “minimize losses” move. Cancel Fuchsia R&D, slap Android on the existing hardware, sell it, and cancel the smart speaker product line down the road. Looking at Youtube’s war on ad-blockers starting up, Google management has really started penny pinching. Fucshia is probably still in development, but smart speakers are probably getting cut.
dark2,
If that’s what they want to do, then so be it. Your argument that they’re not doing it for business reasons makes sense. Also, anevilyak’s argument that it might be due to infighting at google makes sense too. But the argument that they’re not doing it because of crypto requirements makes very little sense to me.
I’d expect it has more (or at least a significant degree) to do with Google’s apparent loss of interest in Assistant, in favour of their “new-shiny-thing” vague, half-baked A.I. ambitions:
https://arstechnica.com/google/2023/05/google-at-i-o-2023-weve-been-doing-ai-since-before-it-was-cool/
Not to mention the rapidly declining enthusiasm for smart speakers in general, at least from most of the companies making them:
https://arstechnica.com/gadgets/2022/11/amazon-alexa-is-a-colossal-failure-on-pace-to-lose-10-billion-this-year/
Given their track record, and the recent indications that Google’s disorganization/lack of focus hasn’t improved (if anything, it seems to steadily be getting *worse*), at this point I expect Assistant is going to be seeing the same sort of treatment as Google’s various redundant chat services. Namely, we’ll end up with 3 or 4 competing products from the same company, which manage to both overlap/duplicate each other’s functionality – while also all missing one or two basic features that are present in one or more of their other offerings… before eventually being unceremoniously killed off, in 3 or 4 years, after they’ve stopped updating it for a few quarters, and that (along their hopeless mismanagement of it) causes the usage numbers far enough that they can justify it.
It may sound crazy, but maybe Google is running out of money.
They have only one product, that turns a profit (ads).
And that one has been in free fall for the past 1-2 years.
I have heard numbers ranging from -40% up to -80%.
And (more or less) coincidentaly they started to push hard for youtube premium and against adblockers.
smashIt,
I noticed that too, but so far the adblockers still seem to be winning. Maybe this is why google are testing browser attestation (ie DRM) for chrome. It would give google a new tool for blocking modified clients.
What is the debate? If Fuchsia is still relevant to Google? or is it just not going to be available for some Nest smartspeakers for it’s chipset limitation?
Being an OS aficionado, I will like to see Fuchsia OS running in a way I can play with it on a PC. I personally want to see the open source microkernel (Zircon is MIT license AFAIK) that can compete with the Linux Kernel.
I’m in the same boat. Really looking forward to an open source microkernel compatible with x86-64.
QNX looked really promising but the parent company abandoned the x86 QNX project.
The bad news: OS diversity will remain low, with the Linux, Windows NT, and XNU kernels accounting for say 99% of devices.
The good news: Google won’t replace GPL-licensed Android with their BSD/MIT/Apache-licensed Fuchsia. Using a permissive license is the first step to going proprietary. I doubt that Fuchsia ROMs would have been possible in the same way as Android ROMs (Graphene, Lineage, Calyx, Divest), because I’m sure Google would have kept core parts of the system proprietary, beyond just Google Play Services, even if Fuchsia is nominally open source.
The very good news: Google is still not “reliable” and they’re killing all their products that are not Search/Chrome/Gmail/Maps/Android/Youtube. I don’t think there’s been a single Google product that came out in the last 5-10 years that’s been successful or not killed by Google again. I hope that eventually this will lead to the demise of Google and them becoming a dinosaur like IBM, still making cash but with less soft power.