Google introduced the Tensor series of chips with the Pixel 6 series in 2021. It was meant to usher in a range of chips specially tailored by Google to keep up with its class-leading AI-based software features. Just three Tensor chips in, and many of Google’s new AI features need to be off-loaded to...
They where on Nexus for a while before that too. Now if we could get one that doesn't constantly scan virtually all activities for 'an improved user experience and personalization of the web' out of the box that'd be a pretty awesome next step.
They're claiming that after the Tensor 4 design that Google will be moving to a 3nm TSMC fully custom design.
So the "life support" comment apparently applies to just their current Exynos-derived designs that will need to continue shipping until then, despite the designs being dated (i.e., on "life support").
Not for long with Google cranking up the ads (amount of and length) as well as them stepping up their ad-blocking detection. Newpipe and such still work, but for how long?
I think the problem is the conflicting goals that Google has with that chip. They want the chip to be able to run AI stuff locally with the Edge TPU ASIC that it includes, but at the same time Google also wants to make money by having Pixel devices offload AI tasks to the cloud. Google can't reconcile these two goals.
I don't think they're opposing goals. Google does not make more money from a task running in its cloud than on its devices, if anything that costs them more money.
I think it's realistic to assume that Google is going to impose quotas on those "free" AI features that are running on the cloud right now and have people pay for more quota. It makes no economic sense for Google to keep offering those compute services for free. Remember Google Colab? Started completely free with V100 and A100 GPUs, now you have to pay to just keep using a simple T4 GPU without interruption.
There's a solution: Charge the customer once for the hardware and then add a monthly fee to be able to use all of it. Sony and Microsoft have great success with that.
I worked for abbot labs and we made certain chemicals but we did not discount our product for interior sales so we usually bought the competitor which we could get for a bit cheaper. Makes no sense to me that they would not sell internally wholesale.
All big tech should just be broken into parts. Glass-Steagall the fuckers. Google can have their adtech empire, strip everything else away from them though. Same with Microsoft. Fuck having like 18 departments under a massive main brand, break down the departments into their own corps.
I think in general, having NPUs on devices is such an underrated value of which Google's TPU would be classified as. There's actually a lot you can use them for. The main thing for me personally is definitely voice detection stuff. Although I have to admit FUTO's voice detect which uses whisper, really great. This reply is being crafted entirely using it.
Background noise removal which definitely helps with Speech detection. So it's really nice to have that too. And of course you have things like video processing and everything else Google is bragging about.
I don't think these devices are incapable of doing what Google wants them to. I think it's a mixture of they're simply not fast enough to do it real time, which Google needs for their premium feeling this as well as just not wanting to invest the time in it.
It is interesting that the price has already dropped on the P8 and is expected to continue to drop thru black Friday.
This is precisely why I never buy new anymore, having been burned by two previous Pixel releases in the past.
This chip situation doesn't bode well for my continued Pixel use however.
Software is important but it isn't everything, and like the article said raw horsepower does matter. For me. The most important things are battery, life display brightness, and cellular connectivity - something my pixel 6 Pro objectively fails at on all three fronts.
Combine that with all of the data theft that Google software utilizes, and I think I'm pretty much done.