I've heard of it, but hesitant to buy and find it doesn't work. I've seen on some forums that it doesn't work with the newer macs (e.g. M4, M5), but it still seems to work with the Neo.
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
macOS 27 Will Mark the End of an Era
- Thread starter MacRumors
- Start date
- Sort by reaction score
I know some of them do (Surface laptops have been unusable for me for a while now), but it's possible to buy laptops that don't use it. I'm typing this on a Dynabook X40L-M which has an EyeSafe display and an Intel Ultra 7 processor and it definitely does not do dithering.
Portégé X40L-M Laptops | X Series Laptop | Thin and Light Laptops
Explore Dynabook's innovative laptop computers, notebooks, netbooks and PC accessories. Find the best fit for your home or business with reviews, features, research, information and deals at us.dynabook.com.
us.dynabook.com
Last edited:
Skwoodge
macrumors regular
The writing's been on the wall for a while now. I think this be might a reason for the "Snow Leopard" - esque optimization of macOS this year. They're trimming the fat. And now with the Neo and macOS running on a mobile chip with 8 GB of RAM, macOS has to be streamlined so it doesn't slow it down. So I think every Apple Silicon Mac is going to get a fairly significant performance boost this year.
Mousse
macrumors 601
boppin
Suspended
Huh? There are still people who "only" have 8 GB of RAM in their "old" MacBook Air and Pro, Apple Silicon, and they're getting along just fine with it.
Skwoodge
macrumors regular
Did I say they didn't? This is such a strange post. I'm saying this will benefit other Macs ... especially Macs with less RAM ... because they can't cripple the Neo with the next macOS. You just weirdly invented something I didn't say so you could attack me on it. I have an M1 MacBook Air still and I'm looking forward to seeing how this new streamlined macOS breathes some new life into my Mac. Just like AI finally forced Apple to include more base RAM, the Neo is going to force Apple to optimize macOS more for older hardware.
I’ll be awaiting your report-out on day 1 of the system file size and if you plan to jump ship if it’s 1MB higher 😂
boppin
Suspended
I didn't attack anyone. Maybe you should take a break from the internet.
Skwoodge
macrumors regular
You quoted me and then replied to something I never said ... and now you're being hostile to me pointing that out.
Indeed. My M1-chip MacBook Air is still a pretty fast workhorse, especially with the 8-core graphics and 16 GB of RAM.
As for dropping Intel support, I don't mind too much, since said M1 Air is pretty much my main computer, with me having an 2019 16" MacBook Pro with 6-core Intel i7 processor and 32 GB of RAM as a "backup" computer, largely for the few things I do that may still require an Intel-based Mac...
It still boggles my mind how my M1 MacBook Air even manages to slightly outperform that Intel Core i7 MacBook Pro that's just a year older.
It's an OS level app, the CPU/SOC shouldn't have an effect on it. However it obviously can't remove built in panel dithering. I understand the hesitation of course, but coming from someone with an m4 era device, it works great.
v0lume4
macrumors 68030
I don't see enough people talking about the impacts of Rosetta being sunsetted. There are many apps, some of which well known ones, running on Rosetta. If those don't get a native ARM version, they're toast.
And what of CrossOver? Doesn't it rely on Rosetta to get a lot of apps working?
And what of CrossOver? Doesn't it rely on Rosetta to get a lot of apps working?
LFO8
macrumors 6502
Honestly, I'll be glad to see it go. We are six years down the line here now and some software developers have dragged their feet up until this very very last moment to finally make their **** fully Apple Silicon native. Waited until the very very last moment until their had is forced because they didn't want to allocate the resources or whatever..
I jumped into the AS game in 2024 and never installed Rosetta. I just dropped the companies that didn't have their affairs in order.
I agree. As a matter of fact, devs got two more years this time around than they did in the PowerPC to Intel transition.
Skwoodge
macrumors regular
Don't worry ... he has bloated and buggy Windows (to the point that Microsoft had to make a public apology) that now doubles as spyware (they aren't sorry for ... i.e. Browsergate) and Linux that barely has any app support from major software companies. Wonderful alternatives out there.
indefinite
macrumors regular
Regarding your 2017 27" iMac, you could perhaps try installing Sequoia (via OCLP) on an external SSD and booting up from the external drive. There's no risk in testing this out because Ventura remains installed on the iMac's internal drive. I did exactly this on my 2017 27" iMac 5K (32GB RAM) and have been quite satisfied with its overall performance.
The only significant glitch I've encountered so far is that my iMac will often not wake up from sleep properly when booted from the external Sequoia SSD, requiring a hard restart. So I now just shut down my iMac whenever I am finished with a task rather than letting it go into sleep mode. That may be not be an acceptable work around for everyone but it's been satisfactory for me, and I rarely boot up from my iMac's internal Ventura drive anymore.
Anyway, there's no harm in trying, especially if you have a spare SSD lying around.
Intel was a huge step forward when it was introduced, but over time Intel just couldn’t keep up with what Apple needed. Once iPhones launched with ARM processors, the writing was on the wall. By the end, Intel chips were the Achilles heel of so many otherwise great Macs.
But let's not totally write off dual boot into Windows. I think a lot of people found that super useful for a while (and perhaps still).
It is a big change if you use one of those Macs and plenty of regular users do.
Yeah, I still do; I have en external nVidia card which does not play nice with macOS - but happily runs diffusion models and LLMs from Windows. Also, gaming. And I need to run Windows VMs for a fair part of my work; e.g. PowerBI, the occasional Report Builder (yes, I am apparently a masochist) etc. Dunno how well that would fly on an ARM.
Anyway, I am not going to be missing OS updates for that machine terribly, and security updates will still be pushed out for the foreseeable future 🙂
mansplains
macrumors 68000
Docks, SSDs, drive enclosures. They all work fine so I'm not planning to spend money to replace them AND carefully check if they will update with ARM Windows or Linux.
An x86-64 Mac covers 100% of my use cases and ARM Macs don't . That's ultimately what matters to me and I'm not trying to convince others that they're wrong.
dustinisatexan
macrumors member
You’re the one that doubled down on incorrect information when they corrected you, instead of just admitting you were wrong in the first place… look in the mirror instead of clutching to your own post maybe?
Gioser
macrumors member
With the progressive phase-out of Intel-based Macs, macOS can evolve into a fully ARM-native operating system, eliminating the need to maintain compatibility with the x86_64 ecosystem. This allows Apple to remove entire layers of abstraction and legacy support, including binary translation (e.g., Rosetta), legacy libraries, and multi-vendor driver stacks (AMD/NVIDIA/Intel GPUs) that were historically required to support heterogeneous hardware configurations.
The result is not just reduced codebase complexity, but also improved system predictability, tighter performance control, and a more streamlined development pipeline across the compiler, scheduler, and memory management subsystems.
Additionally, the architectural convergence across macOS, iOS, and iPadOS (all ARM-based) removes the need for cross-ISA portability. Applications can now target a unified runtime and instruction set, without maintaining separate code paths or dealing with x86-to-ARM translation. This significantly lowers development overhead and improves software efficiency across the entire platform.
On the AI side, the distinction is even more pronounced.
All Apple Silicon SoCs have included a dedicated Neural Processing Unit (Neural Engine) since 2020, specifically designed to accelerate machine learning inference workloads locally, with high throughput and energy efficiency. This allows both the operating system and applications to offload AI tasks to specialized hardware, avoiding unnecessary load on CPU and GPU resources.
It is also important to clarify a commonly overlooked point: no Intel-based Mac has ever included a dedicated NPU. Intel only introduced CPUs with integrated neural processing units starting in December 2023 (Core Ultra architecture). This means that the entire Intel Mac lineup (including relatively recent models) lacks any form of dedicated AI acceleration hardware.
As a result, all AI workloads on Intel Macs must rely on general-purpose CPU and/or GPU execution, which is inherently less efficient in terms of performance per watt and latency compared to dedicated NPUs.
More broadly, within the x86 ecosystem, NPU adoption is very recent: AMD introduced its first AI-capable processors (Ryzen AI) around mid-2023, while Intel followed at the end of 2023. This means that a large portion of the currently deployed hardware base is not optimized for local AI workloads.
In summary, Apple’s advantage here is not ideological but architectural: a fully aligned ARM-based ecosystem, with vertically integrated hardware and software, and widespread availability of dedicated neural acceleration for several years. This enables more efficient, scalable, and consistent implementation of on-device AI features compared to a fragmented and transitional x86 landscape.
The result is not just reduced codebase complexity, but also improved system predictability, tighter performance control, and a more streamlined development pipeline across the compiler, scheduler, and memory management subsystems.
Additionally, the architectural convergence across macOS, iOS, and iPadOS (all ARM-based) removes the need for cross-ISA portability. Applications can now target a unified runtime and instruction set, without maintaining separate code paths or dealing with x86-to-ARM translation. This significantly lowers development overhead and improves software efficiency across the entire platform.
On the AI side, the distinction is even more pronounced.
All Apple Silicon SoCs have included a dedicated Neural Processing Unit (Neural Engine) since 2020, specifically designed to accelerate machine learning inference workloads locally, with high throughput and energy efficiency. This allows both the operating system and applications to offload AI tasks to specialized hardware, avoiding unnecessary load on CPU and GPU resources.
It is also important to clarify a commonly overlooked point: no Intel-based Mac has ever included a dedicated NPU. Intel only introduced CPUs with integrated neural processing units starting in December 2023 (Core Ultra architecture). This means that the entire Intel Mac lineup (including relatively recent models) lacks any form of dedicated AI acceleration hardware.
As a result, all AI workloads on Intel Macs must rely on general-purpose CPU and/or GPU execution, which is inherently less efficient in terms of performance per watt and latency compared to dedicated NPUs.
More broadly, within the x86 ecosystem, NPU adoption is very recent: AMD introduced its first AI-capable processors (Ryzen AI) around mid-2023, while Intel followed at the end of 2023. This means that a large portion of the currently deployed hardware base is not optimized for local AI workloads.
In summary, Apple’s advantage here is not ideological but architectural: a fully aligned ARM-based ecosystem, with vertically integrated hardware and software, and widespread availability of dedicated neural acceleration for several years. This enables more efficient, scalable, and consistent implementation of on-device AI features compared to a fragmented and transitional x86 landscape.