NotebookCHECK - Notebook Forum

The author wants all AI tasks processed on device.  This would result is high power consumption, accuracy issues and virtually no AI model training. 
The beauty of a cloud powered chip hybrid is you get all the pros and non of the cons.  Maybe the author camps out in the woods frequently where there is no reception of any kind?

The con of clouds is data abuse.

Quote from: RobertJasiek on November 18, 2023, 20:37:43The con of clouds is data abuse.

The con of on-device chip is being about infinitely slower than cloud processing. I mean, one on-device chip vs thousands of dedicated and specialized AI/ML machines in the cloud, hmmm... Didn't Google buy/get some 20,000 H100 GPUs from Nvidia recently?

• https://nvidianews.nvidia.com/news/google-cloud-and-nvidia-expand-partnership-to-advance-ai-computing-software-and-services
• https://cloud.google.com/blog/products/compute/introducing-a3-supercomputers-with-nvidia-h100-gpus

And one Snapdragon 8 G3 is supposed to outperform 20,000 GPUs, lol? Sure, if Pixel has no network connection at all including 2G.

But the author of this opinion is focused solely on benchmark numbers where more is better, no matter of implementation. So let's just say that absolutely all cars in the world are on life support because prototypes and F1 cars are much faster.

AI on SoC is there to better understand your needs and serve you ads, not to let you use Stable Diffusion. )
Sarcasm, but partially true.

I am not even sure that is sarcasm at all because it really is true in general overall usage.

Gee, do some authors at Notebookchat like beating up on Google much?

I'm sorry but this "article" (and a few others like it) reads more like a 'rant' then anything else.

Given that 'smart phones' have been around long enough now that developers are getting overly 'creative' coming up with new features to highlight their product offerings over other OEM's.

While 'AI' seems to being billed as a 'must have feature' or at least 'trendy feature of the month/year' does every user need or want it?

I think if you looked back at Googles history they have made other 'not so smart' moves in the past. (much like ALL the other OEM's)

The phones are all so capable. That 98% of owners don't even notice the difference between the chips.
We all just Doom Scroll into and out of your article hardly pushing this silicone to its limit.

Frankly, I think this push to do all AI on device is misguided. Look at Siri. It tries to do everything on device and it's one of the most idiotic companions ever. On the pixel 8 pros when they tried to sequester the assistant onto the phone. The assistant is pretty much as stupid as serious now. As evidence of this on my s23 ultra, when I use the Google Assistant it actually gets things done. Why? Because it has to go to Google servers where there's much more information available than can ever be stored on device. So the fact that the tensor G3 is having to upload video boost to Google service for processing. I think it's a good thing. Frankly, I think that Google should stop focusing on trying to run everything on device and refocus on making the modems much more efficient. So you can send stuff to the cloud and come back. For a while they were saying how their assistant could do more and with smarter than Siri. And until they seequestered the assistant onto the pixels. This was true. When the assistant runs better on a third party Android device then on your own device? Because on a third party device it has to come to your servers to get the job done. It should be obvious that your strategy a sequestering the assistant in anything else to on device is simply not working and makes you look no better than Apple does in terms of AI performance.

Samsung made a fool of Google. They crippled the Pixel line of phones, neutering a competitor. At the same time, Samsung extracted huge fees from Google. It should have occured to someone at Google that there was an excellent reason Samsung wasn't pursuing Google's Tensor design in Samsung's own phones. After billions wasted the Tensor line has proven no better than midrange Qualcomm processors. I must recharge my Pixel 7 twice to make it through a full day. Samsung uses Qualcomm processors, so should Google.

Just just MediaTek chips Google

And yet, my 6a is running just fine. Stop with your Apple shilling bullshit.

So funny to see Google fanbois defending the pathetic Tensor under lacklustre Pichai's watch. Enjoy your Tensors and Google marketing lies.

well. pixel biggest problem today is that it's tensor are so weak compared to competitors. if the solution is ends up in the cloud, then better just use your competitors chip, especially when after three iterations, they still can't make it right. it still hot, throttling and underperforming.

Damn this article reads like an a.i. generated post.

If someone does not like what the author described, it does not mean that it is not true.

The fact is that mobile chips can handle offline requests well. An example of this is HiSilicon chips from a few years ago.

So this goal - to do what others have already done several years ago, is not superfluous. The question is to approach its implementation as thoroughly as possible.

Because even Samsung, which does not make the best chips, still makes normal devices, because it controls many processes. Therefore, Google should approach the production of smartphones not in the "we will tell them - they will make it for us, and then we will collect everything together" type, but really involve its own experts in this, immerse themselves in it as much as possible, bring the optimization of the OS itself to order [and not hope, since it already exists, it seems that it should work somehow].

Why does this article read like a teenage boy saying everything bad he can think of about the girl who just turned him down for a date?

Quote from: Damacus on November 19, 2023, 07:20:52So funny to see Google fanbois defending the pathetic Tensor under lacklustre Pichai's watch. Enjoy your Tensors and Google marketing lies.

Not all of them. Please do not lump us all in the same category. Many of us refuse to touch/upgrade to any Exynos-based Tensor chip Pixels with a 10 ft bargepole. Yes, we are well-aware it is complete trash. Also doubt these are actual true fans at all because if they were they'd be more critical on the company they're spending money on.

Quote from: Will on November 19, 2023, 04:55:37And yet, my 6a is running just fine. Stop with your Apple shilling bullshit.

Uhh, by the same token even a 7 year old phone is "just fine" if all you do is basic messaging, email, surf and youtube. Why not just buy a old used phones if not interested in actual progress? This has nothing to do with Apple shilling. Nobody should be defending this sorry excuse of a chip, AI/fanboism-shilling or not, it's inexcusable.

Interesting points raised, but there are likely many factors not considered in the conclusions drawn.

More questions should be asked than statements made, as a rule of thumb for a modest appraisal of the situation.

Have you considered the price comparisons (including launch bundles), profit margins, long term goals, marketing, etc.?

One possible problem may be capabilities rather than speed.  Some of the AI/ML solutions (like the one in my 11th gen Intel CPU -- which Intel seems to be preternding doesn't exist since they've commented on the 13th or 14th gen being the first to have ML acceleration in it)... the capabilities are very limited.  CUDA it ain't.  Very limited data set size, only supporting 1D input (i.e. audio, not even a 2D table of data).  My understanding is in Windows it's used for microphone background noise removal. In Linux there's a driver for it and I got it to show up in Tensorflow but it was too limited to run anything that basically wasn't specifically designed with it's limitations in mind so I got no "out of the box" models to train or predict with it.

I don't know what the Google Tensor's capabilities are but it could be it's missing FP32 or FP64 support, data size or shape limits, etc. make it unsuitable for running these things Google is running in cloud.

You murdered Andrei Frumusanu's name. 🤦🏻�♂️

What is the author talking about??

Suggesting that a mobile device should be able to run those large generative AI models?? That's ridiculous. Show me a Qualcomm chip that can do that!

Also saying that processing on-device will take less power... Have you ever precessed video? It's a much heavier task than networking.

With that said, yes, the Tensor chips are not great. Mostly due to heating up, in my opinion. I do not care about high-end performance, but about efficiency - YES, I DO.

Quote from: nokite on November 20, 2023, 19:54:40What is the author talking about??

Looking at benchmark numbers where more is better and doing a whole opinion based on that, lol.

Quote from: nokite on November 20, 2023, 19:54:40Suggesting that a mobile device should be able to run those large generative AI models??

Yes 😁

Quote from: nokite on November 20, 2023, 19:54:40Also saying that processing on-device will take less power... Have you ever precessed video? It's a much heavier task than networking.

Very true (also a lot of thermal throttling in many phones).

Well there are native implementations of Stable Diffusion generators running on iPhones' neural engine and GPUs, but it's so useless - slow and hot.

Quote from: Deng Li on November 19, 2023, 03:16:31I must recharge my Pixel 7 twice to make it through a full day.

I don't know, maybe just use GrapheneOS?

Obviously, our usage patterns will differ, but the battery life on my Pixel 7 Pro with GrapheneOS is fantastic.

Quote from: SteveT on November 23, 2023, 09:17:12Obviously, our usage patterns will differ, but the battery life on my Pixel 7 Pro with GrapheneOS is fantastic.

Can't go against physics, SoC has a specific numbers of consumption and a specific battery capacity in Wh. If it makes Pixel fantastic, any other phone with 2x battery life will become, idk, marvellous. But as it's limited to google phones, you simply can't compare.

I don't understand all the moaning in the comments.
It's quite obvious that Tensor is far behind. My Pixel 8 is barely faster than my old OnePlus 8 was, and that's with a 4 year old Snapdragon 865. What's more, it gets hot just messaging people and listening to music, because the modem is inefficient as all hell while also providing terrible cellular reception. Light tasks drain the battery, and it just stops charging if you're doing anything because the battery hits 40°C in no time. You have to recharge within the day if you plan on more than 5 hours of screen-on time.

This chip is supposed to be flagship-class, but it unintentionally continues the Pixel 5's expensive phone with mid-range performance tradition.