Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
The M-series MacBook Pros are so thick it's almost comical. I remember the first time I saw one I couldn't believe I was actually looking at the latest Mac laptop. They're as thick as the MacBooks Apple made in the early 2000s.
That is true for the newer MacBook Pros. I was shocked that the M1 MacBook Pro 16” is quite a bit chunkier than my work’s MacBook Pro 16” Intel model from 2019. Noticeably so.
But it’s not like I travel with it, so moving it from the desk to sofa to bed isn’t much of a strain. The screen is noticeably better on the newer models.
I guess if you want lighter/thinner than there is always the Air.
 
Last edited:
  • Like
Reactions: wildaman
That is true for the newer MacBook Pros. I was shocked that the M1 MacBook Pro 16” is quite a bit chunkier than my work’s MacBook Pro 16” Intel model from 2019. Noticeably so.
But it’s not like I travel with it, so moving it from the desk to sofa to bed isn’t much of a strain. The screen is noticeably better on the newer models.
I guess if you want lighter/thinner than there is always the Air.
Yeah, I understand why they are thicker. And for people who actually need a MBP for work (audio engineers, video experts, graphics people, etc) the thickness of the new Apple Silicon MBPs wouldn't put me off. They are designed for performance – who cares how thick they are if you need one for work?

But like you, I was shocked at just how much thicker they look than the 16" Intel MBP (which I own). Normally I've always bought the big screen MBP whenever a new design came out, but I passed on the Apple Silicon MBP because my Intel MBP looks so slim and sexy compared to it.

For me, all I care about is having a big screen Mac laptop – but I want the body as thin as possible. I don't actually need the power of a MBP for my work. That's why I'm hoping the rumors of a 15" MacBook Air are true.
 
It is frustrating that even experts like this don't really know how to answer the SSD speed issue. Saying, "Well I got the 1tb version that has super fast SSDs, so I have no idea how the smaller SSDs will function." is kind of lame.

Here is how I would explain it. There are two ways your SSD gets used: (1) memory swap and (2) reading/writing files. If you don't regularly read/write multi-GB files, then base model with 256gb on one Nand SSD is going to fine for you on point 2. However, your Mac probably often uses the SSD for memory swap. Your Mac uses the memory swap because it has run out of RAM. If you only have 8gb of RAM, then your Mac is going to be constantly using memory swap to the SSD. This is in ordinary use cases (which I will define as 10+ browser tabs open and 5+ productivity apps open). If any of these apps are Microsoft or Google Chrome, the products will be seriously using RAM at all times. So if you have only the 256gb with one Nand SSD both the non-stop memory swap will be going on when you decide to write/read the SSD. This will create a logjam even though that SSD is very fast. This the issue with the base model M2 mini. In ordinary use, the 8gb is going to be insufficient, the mini will use memory swap, then since the SSD is being used for that, when you go to read/write some large file, you will see some slow down. You can solve this issue temporarily by one simple trick: quit your web browser. But this is annoying. Fundamentally though it is not an issue with the Nand SSD, it is an issue with the Mac not having enough RAM and thereby accessing the SSD for memory swap.

There seem to be three SSD upgrades in these latest models. 256gb is one Nand SSD. Going to 512gb gets you two of the same 256gb Nand SSDs, so two routes to access the SSDs both for memory swap and for write/read for files. Apparently if you go to 1tb you get two 512gb Nand SSDs and these SSDs are twice as fast as the 256gb Nand SSDs that are found in the two lower configurations. So two routes and double speed SSDs on those routes.

If you start with the base M2 mini, either upgrade path will solve the SSD slow down issue. If you get 16gb of RAM, then your mini won't need to memory swap as much. Without being taxed by memory swap, the 256 SSD is going to be plenty fast enough. And a user who is getting a 256gb SSD is not going to be moving multi-GB files around throughout the course of the day anyway. But if you get the 512gb SSD, but keep it at 8gb RAM, then you have two routes to access the SSD. You will still be doing memory swap pretty much all the time. But you can think of it like this: one SSD can handle the memory swap while the other gets the read/write for whatever large file you are working on. Basically at 8gb RAM, your computer won't function at that level without memory swap, so it benefits from that memory swap being faster.

Personally, I would get the 16gb of RAM. I don't work with large files (like I probably don't have a file on my computer that is larger than 100mb (setting aside applications)) and when I do work with my larger files, I work with them one at a time (i.e., I open the file, I save the file). But I do expect to have tons of apps and browser tabs open at all times. What my computer needs is RAM and reducing how much it goes to memory swap and how much it relies on that memory swap is going to be more useful.
 
It is frustrating that even experts like this don't really know how to answer the SSD speed issue. Saying, "Well I got the 1tb version that has super fast SSDs, so I have no idea how the smaller SSDs will function." is kind of lame.

Here is how I would explain it. There are two ways your SSD gets used: (1) memory swap and (2) reading/writing files. If you don't regularly read/write multi-GB files, then base model with 256gb on one Nand SSD is going to fine for you on point 2. However, your Mac probably often uses the SSD for memory swap. Your Mac uses the memory swap because it has run out of RAM. If you only have 8gb of RAM, then your Mac is going to be constantly using memory swap to the SSD. This is in ordinary use cases (which I will define as 10+ browser tabs open and 5+ productivity apps open). If any of these apps are Microsoft or Google Chrome, the products will be seriously using RAM at all times. So if you have only the 256gb with one Nand SSD both the non-stop memory swap will be going on when you decide to write/read the SSD. This will create a logjam even though that SSD is very fast. This the issue with the base model M2 mini. In ordinary use, the 8gb is going to be insufficient, the mini will use memory swap, then since the SSD is being used for that, when you go to read/write some large file, you will see some slow down. You can solve this issue temporarily by one simple trick: quit your web browser. But this is annoying. Fundamentally though it is not an issue with the Nand SSD, it is an issue with the Mac not having enough RAM and thereby accessing the SSD for memory swap.

There seem to be three SSD upgrades in these latest models. 256gb is one Nand SSD. Going to 512gb gets you two of the same 256gb Nand SSDs, so two routes to access the SSDs both for memory swap and for write/read for files. Apparently if you go to 1tb you get two 512gb Nand SSDs and these SSDs are twice as fast as the 256gb Nand SSDs that are found in the two lower configurations. So two routes and double speed SSDs on those routes.

If you start with the base M2 mini, either upgrade path will solve the SSD slow down issue. If you get 16gb of RAM, then your mini won't need to memory swap as much. Without being taxed by memory swap, the 256 SSD is going to be plenty fast enough. And a user who is getting a 256gb SSD is not going to be moving multi-GB files around throughout the course of the day anyway. But if you get the 512gb SSD, but keep it at 8gb RAM, then you have two routes to access the SSD. You will still be doing memory swap pretty much all the time. But you can think of it like this: one SSD can handle the memory swap while the other gets the read/write for whatever large file you are working on. Basically at 8gb RAM, your computer won't function at that level without memory swap, so it benefits from that memory swap being faster.

Personally, I would get the 16gb of RAM. I don't work with large files (like I probably don't have a file on my computer that is larger than 100mb (setting aside applications)) and when I do work with my larger files, I work with them one at a time (i.e., I open the file, I save the file). But I do expect to have tons of apps and browser tabs open at all times. What my computer needs is RAM and reducing how much it goes to memory swap and how much it relies on that memory swap is going to be more useful.
Well said, except for ignoring the absolutely critical fact that a new box not even ordered yet will not spend its 3-6 year working life with the apps and OS that were being used in 2022. Forty years of PC history unequivocally state that 2024-2027 app/OS versions will have evolved to take advantage of more RAM than in yesterday's scenario that you describe.

And a good look at Apple's superb Unified Memory Architecture in the M series chips suggests that the demand for additional RAM to help feed modern SoC performance gains will most likely be increasing even faster than history predicts. IMO Apple also telegraphs that to us as they dramatically increase available RAM with M2 boxes (up to 32 GB available on Mac Mini and up to 96 GB available on MBP).

Most users are likely to one way or another be taking advantage of the new things app/OS designers will be bringing to us over the life of any new box, so personally I recommend minimum 32 GB RAM for most new purchasers reading here.

Unless of course someone has a solidly fixed low-RAM-need for the next 6 years, like granny who only plans some email and simple single-tab browsing forever. Or perhaps low budget K-12 edu (do not get me started on how awful it is that the world's richest nation provides third-world-level K-12 edu...) where clever admins will force-fit apps to suit RAM as necessary, or have better budgeting for SSD replacements than they do for new-box RAM.
 
Last edited:
Hard to understand everything you are asking. But yes, buyers wanting to optimize new computer purchases should

- To the extent feasible, buy more RAM than seems appropriate for today. This has been sound procedure for 40 years, because app and OS designers constantly build to improve performance using more and more RAM, a good thing.

- Buy oversize SSDs to facilitate smooth operation, ~ double one's expected need. An SSD or HDD crash is a huge PITA, and spending a little extra to improve longevity is good finances.
Thanks for your reply. You brought to my attention things I hadn’t thought of, like, if you go for it and spend the money to get an oversized SSD that won’t fill to capacity so quickly, it’ll result in fewer Read/Write/Erase cycles, thus extending the lifespan of your drive.

My current practice over the years has been to always to leave 256MB to 500MB to 1TB (as SSD sizes have grown over this time) free, and making sure to offload files to an external drive whenever it approaches those limits. (I know there’s already overprovisioning, but still…) ALSO! More RAM = less Virtual Memory swapping = less SSD use = longer lasting SSD.

Second, the overall theme of this and other articles about issues with Mac SSD speeds seems to be, the more NAND chips an SSD is composed of, the faster it is. So, a 500GB SSD composed of two 256GB chips is faster than an SSD composed of a single 500GB chip, and so on. This reminds me of how RAID 0 (disk striping) works. Do you think it is similar?

I’ve also read that current NAND chips get HOT! As hot as CPUs, often. So, do you suspect that more thermal throttling/performance decrease occurs on 4TB and 8TB MacBook Pros than configs with smaller SSD drives?

(Although, there is a strange article on Ars Technica that says researchers have found that applying heat to NAND Flash cells can drastically extend their lifespan! Huh?!)

Onetime Apple Software Evangelist, Guy Kawasaki, always said, “Overbuy hardware and underbuy software.” He’d seen too many people skimp on hardware power and, instead, waste a ton of money on too many software apps they never ended up needing/using.

I remember, way back, upgrading my Mac from its 80MB internal drive to a Seagate 400MB drive and all my friends saying that I was insane! (Who’s insane now? Hint: they are. I have a lot of insane friends now.)

Am I ever going to wish I hadn’t gotten 96GB of RAM and an 8TB SSD? I’ve owned too many Macs over too many years to know that I won’t. I might someday wish Apple had had even higher configurations at the time I bought it, but I know I won’t ultimately regret buying the highest cofig today. (I’ve “starred” in this episode before…)

To your point about the pain of HDD or SSD crashes, I’m also upgrading my NAS. I could potentially lose part of one day’s work, but not everything.

P.S. I don’t think anyone knows the answers to my Anobit questions except Apple themselves. But $500 Million?!
 
Well said, except for ignoring the absolutely critical fact that a new box not even ordered yet will not spend its 3-6 year working life with the apps and OS that were being used in 2022. Forty years of PC history unequivocally state that 2024-2027 app/OS versions will have evolved to take advantage of more RAM than in yesterday's scenario that you describe.

And a good look at Apple's superb Unified Memory Architecture in the M series chips suggests that the demand for additional RAM to help feed modern SoC performance gains will most likely be increasing even faster than history predicts. IMO Apple also telegraphs that to us as they dramatically increase available RAM with M2 boxes (up to 32 GB available on Mac Mini and up to 96 GB available on MBP).

Most users are likely to one way or another be taking advantage of the new things app/OS designers will be bringing to us over the life of any new box, so personally I recommend minimum 32 GB RAM for most new purchasers reading here.

Unless of course someone has a solidly fixed low-RAM-need for the next 6 years, like granny who only plans some email and simple single-tab browsing forever. Or perhaps low budget K-12 edu (do not get me started on how awful it is that the world's richest nation provides third-world-level K-12 edu...) where clever admins will force-fit apps to suit RAM as necessary, or have better budgeting for SSD replacements than they do for new-box RAM.
Damn, 32gb huh. You are somewhat preaching to the choir here because I've been on the "you get 16gb or you don't buy" bandwagon for two years now. But the way Apple prices RAM, I would not suggest $400 of RAM upgrades on the lower cost machines. That is just putting too much extra money into the machine. Like I would get an M2 mini with 16gb, but I wouldn't spend another $200 for 24gb of RAM. I would not spend $400 to upgrade the M2 Pro mini to 32gb unless I specifically knew that somehow I had a RAM intensive use case but not a use case that pulls me to the Mac Studio. That is a truly niche set of people though.

I think 16gb is going to be fine for several years and then either (A) you are the granny if you are still using the same machine 5 years from now or (B) you are selling/handing down the machine to granny and it will be fine for the remaining three to five years of useful life with 16gb. Apple is just selling too many 8gb machines now to not maintain a path for 16gb machines over the next decade.

The new pro laptops though are a tougher call. And I suspect that is where you are focused. They are $1,999 to start. So that $400 upgrade is not a big percentage increase in cost. And you may very well be right that 4 to 6 years from now those laptops might struggle at 16gb but be fine if they had 32gb of RAM. [Edit: struggle in high use cases. I think 16gb will be fine in 4 to 6 years for lower level use cases.]
 
Last edited:
Thanks for your reply. You brought to my attention things I hadn’t thought of, like, if you go for it and spend the money to get an oversized SSD that won’t fill to capacity so quickly, it’ll result in fewer Read/Write/Erase cycles, thus extending the lifespan of your drive.

My current practice over the years has been to always to leave 256MB to 500MB to 1TB (as SSD sizes have grown over this time) free, and making sure to offload files to an external drive whenever it approaches those limits. (I know there’s already overprovisioning, but still…) ALSO! More RAM = less Virtual Memory swapping = less SSD use = longer lasting SSD.

Second, the overall theme of this and other articles about issues with Mac SSD speeds seems to be, the more NAND chips an SSD is composed of, the faster it is. So, a 500GB SSD composed of two 256GB chips is faster than an SSD composed of a single 500GB chip, and so on. This reminds me of how RAID 0 (disk striping) works. Do you think it is similar?

I’ve also read that current NAND chips get HOT! As hot as CPUs, often. So, do you suspect that more thermal throttling/performance decrease occurs on 4TB and 8TB MacBook Pros than configs with smaller SSD drives?

(Although, there is a strange article on Ars Technica that says researchers have found that applying heat to NAND Flash cells can drastically extend their lifespan! Huh?!)

Onetime Apple Software Evangelist, Guy Kawasaki, always said, “Overbuy hardware and underbuy software.” He’d seen too many people skimp on hardware power and, instead, waste a ton of money on too many software apps they never ended up needing/using.

I remember, way back, upgrading my Mac from its 80MB internal drive to a Seagate 400MB drive and all my friends saying that I was insane! (Who’s insane now? Hint: they are. I have a lot of insane friends now.)

Am I ever going to wish I hadn’t gotten 96GB of RAM and an 8TB SSD? I’ve owned too many Macs over too many years to know that I won’t. I might someday wish Apple had had even higher configurations at the time I bought it, but I know I won’t ultimately regret buying the highest cofig today. (I’ve “starred” in this episode before…)

To your point about the pain of HDD or SSD crashes, I’m also upgrading my NAS. I could potentially lose part of one day’s work, but not everything.

P.S. I don’t think anyone knows the answers to my Anobit questions except Apple themselves. But $500 Million?!
You ask "This reminds me of how RAID 0 (disk striping) works. Do you think it is similar?"


Yes. RAID 0 striping is exactly what the performance of multiple-SSD operation in Macs reminds me of. My personal opinion is that only a few folks should be buying the smallest SSD offerings or the smallest RAM options anyway. More SSDs in a box may perform better as one reads while another writes or whatever, but also note that larger SSDs are often faster drives.
 
Last edited:
I wonder if Dan hits the nail on the head for a lot of MBP users.

If you never hear the fans are you wasting money on the computer and can get a lower spec one, as all it suggests to me is you don't challenge the hardware.
Nonsense. "Hearing the fans" correlates to a) ambient noise, b) personal hearing sensitivity and c) computer heat production; not to inadequate performance.

E.g. my 2016 MBP is overloaded very frequently (when being used in desktop mode with 3 external displays) but I seldom hear the fans. And sometimes I have specifically noted no fans operation while waiting on the SBBOD.
 
Last edited:
  • Like
Reactions: R2DHue
Nonsense. "Hearing the fans" correlates to a) ambient noise, b) personal hearing sensitivity and c) computer heat production; not to inadequate performance.

E.g. my 2016 MBP is overloaded very frequently (when being used in desktop mode with 3 external displays) but I seldom hear the fans. And sometimes I have specifically noted no fans operation while waiting on the SBBOD.

I could make the 2016 MBP sound like a plane when connected to 4K displays with GPU intensive work. Can‘t hear anything on the new machines unless I stick my ear against it.
 
It sounds like you think that disruptive innovation is something that should be an annual routine. It should not. M2 should be an incremental innovation.
I fear the same. That “M3 will not be an industry ‘disruptive’ chip.” I fear its primary gains will be those that automatically come with the 3nm process: smaller, denser die, more transistors, less heat and lower power consumption and potentially higher clock speed.

But as for CPU and GPU? I fear Apple will be relegated to a “brute force” approach: more of the same, just more.

I fear that we’ll get more of the same “Avalanche” and “Blizzard” cores from the iPhone A15 SoC — but in higher numbers. I.e. fourteen of the same M2 Performance cores; four of the same M2 Efficiency cores. Or maybe 16/6 or 18/8 and so on. Maybe they’ll tinker with the system level cache a bit. Maybe they’ll do more with the secretive AMX “cores.” Maybe we’ll get LPDDR5X RAM for a 1.3x increase in RAM speed. ARMv8.4…

And I fear that the GPU won’t change either, we’ll just get more of them. 24 instead of 19, 48 instead of 38, and so on. Do this all at a higher clock rate and Apple might claim to the world that it’s a “generational change” for the industry.

Fuse four “effective M2s on 3nm” together instead of 2 and call it Ultra.

Apple must have screwed up bigtime to lose so many (apparently unhappy) Apple Silicon Engineers, IDK what Apple did…

It seems like there is a crippling “hole” in Apple’s Silicon Engineering team at the moment. (Hope I’m wrong.)

I think a gaping hole was blown in Apple’s Industry-leading Silicon Engineering team when the Chief Architect of Apple’s Silicon for 10 years, Gerard Williams III, left Apple in 2019 and took quite a few key Apple Engineers with him to start Nuvia. But Apple can’t buy Nuvia like it did Next because 1.) Gerard Williams doesn’t like working at Apple, and 2.) Because Qualcomm has already snatched up Nuvia for a cool $1.2 Billion. Qualcomm has since pre-announced the new (allegedly “built from the ground up”) Oryon CPU core technology.

If true, the problem is, as Apple spends time attracting new talent to replace that which left, and as Apple (hopefully) aggressively poaches Silicon Engineers from other chipmakers, precious time is elapsing. Time its ruthless competitors are using to catch up to and perhaps surpass Apple.

Remember when many articles were written about how far Apple was ahead in its SoC designs and how long it would take for competitors to catch up?

It shook the industry when it was found out that the Apple A7 was 64-bit and was 200% faster than its predecessor in both CPU and graphics performance.

Apple beat ARM to ARMv-8 ISA and 64-bit. Apple was leapfrogging others in the industry and seemed like an unstoppable juggernaut that would be out ahead of everyone else for years.

But then its industry-leapfrogging advancements seem to have stalled. Will Apple Silicon become the next “PowerPC” of the industry as competitors surpass it and leave it in the dust? I sure hope not.

ARM has beaten Apple to ARMv9.0A and ARM has beaten Apple with ray tracing hardware called the ARM Immortalis-G715 GPU core; Samsung has its Exynos 2200 SoC with ray tracing support in hardware out now; Nvidia, and Intel (Arc) and AMD (Xclipse) and MediaTek all have hardware ray tracing in the works.

Qualcomm supposedly has ray tracing technology in hardware that can’t even be carried out through software! (I don’t get it, but…)

And Qualcomm’s Snapdragon 8 Gen 2 SoC — expected in the first half of 2023 — will not only have hardware accelerated ray tracing and DLSS, but one main ARMv9 CPU core in addition to multiple mid-tier and efficiency cores, on-die Natural Language Processing, an integrated Snapdragon X70 modem/radio on the SoC instead of as a separate chip. It will also feature WiFi 7, hardware level Dynamic Spacial Audio processing and dual Bluetooth for greater connectivity (like a better Apple H2 chip integrated onto the SoC/SIP), so-called “A.I assisted” sub-6GHz and mmWave cellular connections, and simultaneous dual-SIM support for switching phone calls between both SIMs at one time.

Unlike Samsung and others, Qualcomm doesn’t really care what customers or platforms it sells its proprietary ARM-based SIP to. Qualcomm doesn’t make phones or computers. So it is expected that Windows will completely migrate to ARM (just like Apple/Mac did) sooner than thought. The upcoming Snapdragon 8 Gen 2 may be key for Microsoft. Or Intel may be key. I would hate if Windows machines eventually compare to Apple Silicon Macs. 😡 Or if industry pundits say so whether they do or not! :mad:

The leaked news that Apple “planned a generational leap for the graphics processor in the latest version of its high-end smartphones, the iPhone 14 Pros” but had to abandon it supposedly for reasons of unacceptable power consumption/battery drain (but who knows?) is a troubling sign for Apple’s silicon prowess.

Apple’s own announcement that it has paused development on its own radio/modem chip in order to concentrate on Apple SIPs is also troubling news for a company so big and capable.

It seems like working for Apple is no longer a “cause” but just a job. It seems like people don’t feel like they’re working to change the world and improve humanity but are working for a NASDAQ ticker symbol instead. “Vision” seems to be lost and Apple seems to be more of a corporation just like any other. That’s why Apple employees leaving to work for Microsoft or Google doesn’t feel like that much of a departure.

(I hope I’m completely wrong about all of this. I’d never be happier to be wrong about anything in my life.)

Qualcomm “stole” — In My Personal Opinion — technology developed at Apple when it acquired Nuvia and all the Apple engineers that came with it. There’s little doubt: a huge chunk of the technology now at Qualcomm was developed in Apple’s labs, on Apple’s dime.

Qualcomm’s public statements are carefully worded to say that the former “Nuvia” started from scratch on the technologies it’s bringing to Qualcomm. (Yeah, right.)

There are lawsuits against Qualcomm over this from both Apple and ARM, but lawsuits “don’t make the donuts.”

Lawsuits aren’t the same as Apple — once again — shocking the entire industry with some entirely new, paradigm-shifting Silicon hardware breakthrough. (And 3nm isn’t it.)
 
Last edited:
Does anyone have experience traveling with their 16, it sounds like it might be a big pain? I don't travel too much once per quarter. But like the idea of larger screen.
I have the M1 max 16 inch, it's a bit heavy but it gets almost everyday with me in my backpack. I was in between the 14 and the 16 and ended up going for the 16inch, never regretted a day.
 
  • Like
Reactions: GCC and kp98077
I have the M1 max 16 inch, it's a bit heavy but it gets almost everyday with me in my backpack. I was in between the 14 and the 16 and ended up going for the 16inch, never regretted a day.
Yes I could see that! I saw many in the airport. iPad for plane is fine for me
 
I've yet to see a single review comparing the ProRes encoding capabilities. Cinebench comes close but not quite. Memory bandwidth and other potential bottlenecks haven't been improved since the M1 Pro and Max but I was hoping ProRes encoding engines had some sort of upgrade, as the engine count per Pro/Max haven't changed (1/2 engines each).

I switched to a noisy PC for encoding purposes because even the Studio's Ultra doesn't cut it, but I could really use a little oomph when I'm in the field and just want to test the output before further shooting. I've a feeling that these new ManBear Pigs still fall short.
 
  • Haha
Reactions: JimmerJabber
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.