danvm

elijahg said:

danvm said:

elijahg said:

loopless said:

It is NOT BS.  Unified memory is a huge advantage.

I have a 16GB 14" M1 MacBook Pro, and a  Dell 32GB Windows 11 Core I7 laptop. Both with SSD's.  I use them for software development.
The Windows 11 machine is bumping up against its memory limits (at which point the performance tanks)  earlier than I have problems with the MacBook when doing a similar set of tasks. For example, using QT Creator and Visual Code, then building large code bases and with lots of other apps open at the same time. 

And lets not talk about the various "blue screens" that still seem to plague Windows.
I looked at upgrading the Dell's memory  but it has CAMM memory that costs $1000 to upgrade - so don't be complaining about Apples prices!

Windows is hideously inefficient with RAM. Doesn't excuse Apple from still only supplying 8GB as standard though. If you need a VM for example,  that will eat all of the 8GB straight up. 

I know that Windows and macOS works differently, but I never seen a test where it shows that Windows is "hideously inefficient with RAM". At least in my customers working with heavy loads, they had no issues at all with memory management in Windows.  But maybe you had a different experience.  

Just anecdotal really - but running the same software for similar time doing a similar thing (Firefox for example), and Windows will have used much more RAM than macOS has. Similarly Windows is always doing something. Even when idle. My Intel MBP running Windows is always hot, same with my work laptop. But the same MBP running macOS at idle is cold. Massive amounts of energy wasted,

Windows has a service, SysMain, that analyze how you use the applications and use RAM as cache to speed up the system.  It's possible that the high RAM usage you see is for cache and not Windows being inefficient.  

In my experience with my Windows devices, there are problems if CPU usage is high while idle.  In normal circumstances, it should be close to 0% of CPU usage.  That's my experience with my Windows devices, and even with my customers. You should check if there are issues with an application or Windows missing an update. 

elijahg said:

loopless said:

It is NOT BS.  Unified memory is a huge advantage.

I have a 16GB 14" M1 MacBook Pro, and a  Dell 32GB Windows 11 Core I7 laptop. Both with SSD's.  I use them for software development.
The Windows 11 machine is bumping up against its memory limits (at which point the performance tanks)  earlier than I have problems with the MacBook when doing a similar set of tasks. For example, using QT Creator and Visual Code, then building large code bases and with lots of other apps open at the same time. 

And lets not talk about the various "blue screens" that still seem to plague Windows.
I looked at upgrading the Dell's memory  but it has CAMM memory that costs $1000 to upgrade - so don't be complaining about Apples prices!

Windows is hideously inefficient with RAM. Doesn't excuse Apple from still only supplying 8GB as standard though. If you need a VM for example,  that will eat all of the 8GB straight up. 

I know that Windows and macOS works differently, but I never seen a test where it shows that Windows is "hideously inefficient with RAM". At least in my customers working with heavy loads, they had no issues at all with memory management in Windows.  But maybe you had a different experience.  

danox said:

danvm said:

9secondkox2 said:

It’s hilarious. Apple has Ben at the forefront of ai for a long time already. 

Being a platform, it makes more sense for them to host generative ai apps. 

But nobody gets that. They actually think small companies like ChatGPT are competitors. 

Yet apple recognizes the future isn’t the status quo. So ChatGPT could make a hardware and platform run if it’s own. So apple heads them off. 

Mits all constant win for apple. Foolish to fault them for building something and only launching when it’s thoroughly ready. 

Apple is historically not first to most technologies after all, just the best. 

Considering that MS is backing up OpenAI and investing billions in them, I wouldn't say that OpenAI is a small competitor.  

Also, I would agree that Apple is not the first and in some cases it's the best.  The problem is that Apple don't have a good record in cloud services.  MS and Google are miles ahead of Apple in this market, even more when you consider the datacenters and infrastructure both companies have.  We'll see what happens in the next few months / years.

Microsoft payed 69 billion dollars for Blizzard, Apple bought for Next Computer for $400 million dollars (which came with Steve Jobs), Apple also spend 750 million dollars for PA Semi, Intrinsity, and Anobit, which led to the A, M, R1 series Apple Silicon SOC's, Apple is not playing catch-up to Microsoft or Google in AI Apple has been extremely efficient in finance and tech execution in comparison to their competition over the years. 

I don't understand how the acquisition of Activision / Blizzard or NeXT Computer has any relation with the discussion about AI.

Apples pathway is on the edge (vertical computer companies can do that) while Microsoft, Google, Meta are designing for the nebulous cloud. Apple has been designing NPUs into their SOCs, to use actual AI-driven features for sometime, in addition Apple has included LiDAR in the Apple Vision which has also been in their iPhones/iPads for the last four years and plays a big part in supplying data for so-called AI functions at the edge.

So Intel and AMD are going to offer CPU based AI/ML acceleration just like the M1, M2 and M3 Macs, iPads, iPhones and even Apple Watches all of which are currently in the hands of the public I hope Intel and AMD and Nvidia can keep the wattage down (doubtful) in their quest to keep up with Apple.

Apple working AI on the edge is not only because they are a vertical company, but also because they don't have an infrastructure to run AI / LMM to the level of Microsoft, Google or Amazon.  And remember that Microsoft is also working in AI on the edge.  

And you are right that Apple have CPU's with neural engines now, but it's possible that the 11 TOPS - 18 TOPS of current Apple M-processors is not enough for what we are seeing today with AI / LMM.  For example, MS will require 40 TOPS to run Copilot locally. That could be the reason there are rumors Apple M4 processors will have a greater focus in AI. If the rumors are true, at the end of this year Apple will have competition from Intel Meteor Lake, AMD Strix Point and Qualcomm Snapdragon X Elite.  I don't think Intel and AMD will match Apple efficiency, but Qualcomm is a different story, considering they have an ARM CPU, their experience with mobile devices and there is a group of people that worked with Apple M processors. We'll have to wait and see how it goes.  Nvidia is a different story, considering they are focused on datacenters.  

9secondkox2 said:

It’s hilarious. Apple has Ben at the forefront of ai for a long time already. 

Being a platform, it makes more sense for them to host generative ai apps. 

But nobody gets that. They actually think small companies like ChatGPT are competitors. 

Yet apple recognizes the future isn’t the status quo. So ChatGPT could make a hardware and platform run if it’s own. So apple heads them off. 

Mits all constant win for apple. Foolish to fault them for building something and only launching when it’s thoroughly ready. 

Apple is historically not first to most technologies after all, just the best. 

Considering that MS is backing up OpenAI and investing billions in them, I wouldn't say that OpenAI is a small competitor.  

Also, I would agree that Apple is not the first and in some cases it's the best.  The problem is that Apple don't have a good record in cloud services.  MS and Google are miles ahead of Apple in this market, even more when you consider the datacenters and infrastructure both companies have.  We'll see what happens in the next few months / years.

danox said:

danvm said:

danox said:

danvm said:

danox said:

danvm said:

danox said:

Apple is hardly behind but they are on a different path and actually are able to execute on hardware and software Apple doesn't need Google the Tensor is 5 or 6 years behind Apple right now and Google is a long way from running AI locally on their feeble Tensor SOC'S and Samsung is (hopelessly) even further away and Microsoft isn't even in the mobile ball game right now, note Wall Street is clueless about this little detail in their AI Hype train.

AI on the edge is the pathway Apple is on, phoning home and having the end user wait is the path Google, Microsoft, Meta, Qualcomm and Samsung are on. Why? Their SOC'S/OS software are way behind Apple. 

What Apple does need to do/consider is leveraging the full benefit of Apple Silicon, the low power/wattage, speed, UMA memory combined with a OS optimized to the hardware just screams out, the M3 Studio Ultra with 256 gigs (coming at WWDC) or a Mac M3 Extreme (CarPlay wasn't all for nothing) Pro tower with 512 gigs just says welcome to the world of inference on the Mac. Apple is/has been working on software tools to support developers for long time.

https://arstechnica.com/gadgets/2020/08/apple-explains-how-it-uses-machine-learning-across-ios-and-soon-macos/.  From 2020 before the current AI Hype

https://venturebeat.com/ai/apple-researchers-achieve-breakthroughs-in-multimodal-ai-as-company-ramps-up-investments/

https://daily.dev/blog/mm1-the-advanced-30b-parameters-multimodal-llm-from-apple

Apple has been consistent in their AI/ML on the edge pathway. Apple won't be using Google anymore for AI than they currently use Google maps.

I think that Apple have no option but AI on the edge.  They don't have the datacenters or infrastructure to run AI / LMM as MS, Google or Meta.  And MS is going for AI on the edge too, with the new processors + NPU from Intel, AMD and Qualcomm.  Looks like MS is the one with all bases covered, from edge to the cloud and AI services.  

No Apple in house data centers yet? Servers are low hanging fruit if that list of companies Google, Microsoft, Meta, Qualcomm and Samsung are involved having a in house OS and Apple Silicon SOC'S are far more important to the future of Apple when it comes to creating new profitable ecosystems.

You forgot Amazon on that list getting rid of Intel was a higher priority to Apple and that modem problem with Qualcomm is still on the higher priority list, after all you can't get to that smaller Apple Vision glasses solution without kicking Qualcomm out.

Google, and Meta want user data that is their sole number one interest in cloud based solutions and Samsung is just along for the ride.

In house OS and ASi are important to Apple since they don't have a data center infrastructure in place for AI / LMM.  That's the reason they are talking with Google for Gemini.  That's different from MS, that is ready to announce AI PC's, so they will have their ecosystem with AI from the edge to the cloud. 

New Era of Work | Home (microsoft.com)
Microsoft to hold a special Windows and Surface AI event in May - The Verge

Will the Windows OS work on day one and will the chip be anywhere within 5 years of Apple, and last but not least will any Windows developers do anything with it. The previous Arm/intel Surface was and is a complete flop Microsoft to this day can't get Windows trackpads to work properly.

We'll have to wait until May to see what MS does with Windows and AI in the MS Build event, and how developers support its features. I don't see why developers will not take advantage of AI / NPU in Windows, as they do in macOS.  
Considering Qualcomm have a group of people that worked in the Apple Silicon design group, I can expect the will have a competitive SoC in the next 5 years. And I'm sure AMD and Intel will be pushing for improving their technology too.

And while I agree that the Surface with ARM was a failure, you have to consider, they are not the only Windows device in the market.  Dell, HP and Lenovo sell a lot more Windows devices than MS itself, and they already announce AI PCs with processors + NPU from AMD and Intel.  Later in the year, we'll see devices with Qualcomm processors. 

https://www.dell.com/en-us/blog/dell-technologies-announces-new-latitude-ai-pcs/
https://news.lenovo.com/pressroom/press-releases/ai-pc-thinkbook-laptops-thinkcentre-neo-desktops-productive-creative-power/
https://press.hp.com/us/en/press-releases/2024/hp-unveils-largest-portfolio-ai-pc.html

Regarding your comment on Surface trackpads, have you use them?  They aren't as bad as you think, even though I would agree Apple does better trackpads.  But since the article is about AI, what do you think of Siri?  Apple have it for 14 years, and still terrible. And now they need help from Google to improve their AI / LMM business. Looks like all companies have something to improve, don't you think?

Google is the one paying Apple billions for a default position so I don't think Apple needs them for anything, and Apple as a ve\rtical computer company definitely doesn't need Microsoft for anything either Apple very wisely passed on buying that fiasco search engine Microsoft was trying to offload a few years ago. Life on the edge (practical real world solutions) is the better pathway for Apple. The Apple Silicon M series, LiDAR and R1 chip lead the way.

After the nebulous hallucinating hot air software on AI dies down Microsoft and Google will be standing around with nothing concrete to show for it like Crypto and Bitcoin which was the last hype of nothing.

Google pays Apple billions for search, not AI / LMM.  From what I know, there are no details on how Apple and Google will work the business relation around Gemini.  We'll have to wait and see. 

And I could agree that Apple was right on passing MS Bing, but I think Apple needs a lot from MS, especially in the business / enterprise market.

And again, AI in the edge is better for Apple because they don't have the infrastructure companies like MS, Google, Meta and Amazon have to run AI / LMM.  This link has a list of the top customer of Nvidia H100 GPU,

The GPU haves and have-nots. - The Verge

If you noticed, Apple is not listed.  And I haven't read that Apple is building datacenters around M-processors.  Maybe they are working in expanding their datacenters and haven't announced it. But based in the article, looks like they had no option but go with a major AI / LMM service like Google Gemini, the same way they depend for cloud services with GCP.

Apple is now Google's largest corporate customer for cloud storage | AppleInsider

That's different from MS, Google, Amazon or Meta, that do not depend on other cloud providers for their AI / LMM services.  

It's true that AI could die in the future, or it cloud be the future, who knows.  Same as the Apple Vision Pro, it could die in the future, or it could be the future.  We'll have to wait and see what happens.