Every computer company charges more for extra features, but Apple has created the perfect pricing scheme with the new M3 Macbook Pros to fool all of us into paying even more for a new laptop.
As always, Apple's entire event was filled with vague performance charts to make the new laptops look as impressive as possible. What's odd is that although most of the charts mentioned the previous M2 chips, every presenter only compared the M3 to the two-year-old M1 series. Maybe they assume M2 buyers won't be upgrading, or maybe the comparisons just sound better.
The $1,600 entry point for the new MacBook Pro with the basic "M3" chip appears to be a decent value, aside from the abnormally low 8 GB of RAM. In my opinion, every laptop over $1,000 should come with at least 16 GB of RAM these days, and being that Apple's RAM and storage options are not user upgradable, almost everyone is going to pay to upgrade them from the factory.
The simple $200 upgrade to 16 GB of RAM will take the average buyer on a wild ride of endless potential upgrades. You might think, "If I'm going to pay for extra RAM, I may as well spend another $200 and get the M3 Pro chip."
This year, Apple boasted about the power efficiency of the new M3 Pro chips, but with further investigation, we learned that Apple seems to have purposefully limited the power of the M3 Pro chips to push users to buy the more expensive Max variant.
The memory bandwidth of the M2 Pro was 200 GB/s, while the M3 Pro is just 150 MB/s. Although the M2 Pro and M3 Pro chips both have 12 CPU cores, the M2 Pro had 8 performance cores and 4 efficiency cores, while the M3 Pro has 6 performance cores and 6 efficiency cores.
The difference between the M1 Pro and M2 Pro and their "Max" variants came down to additional graphics cores. This year, the Max variant also has more CPU cores. So, for someone like m,e who does not require a ton of graphics performance, but who does want the fastest processor, I needed to buy the most expensive M3 Max chip.
As I was building my laptop, I ended up "paying just a little bit more" until my laptop was $4,300, officially the most expensive computer I've ever purchased. But, I can only blame Apple so much. I know exactly what they are doing, and they priced their laptops perfectly this year to pull an extra $1,300 out of me.
My M1 Macbook Pro has been the greatest computer I've ever owned, and I have high hopes for the new M3 Max. For what I just spent, it better be good.
Alternate headline: "I lack self-control and the ability to optimize cost-value tradeoffs."
My good friend Gordon is back!
To be fair, it is tough to predict what you will "need" as the Apple performance benchmarks are vague, at best and even if not, predicting how apps will end up using those resources is also quite tricky and since you can't upgrade later you have to take your shot and hope for the best.
For example, last year, I went all out and went with an M2 Max and went to 96gb of ram. I figured it was overkill, but I wanted the machine to last a long time but there were multiple times this year when working with a heavy AI or special effects workflow I was up around 90gb of usage, so I guess I needed it, but I think that this also because specific applications in my workflow are not efficient with their ram usage. I feel like just as easily a few updates to some ram hungry apps and suddenly that same workflow would be using half the ram.
Predicting is difficult, that's for sure.
Regarding the RAM usage, more efficient memory models will try to use a good chunk of the RAM available - whether it be an app or the operating system. The thought being, if you paid for 96GB of RAM, did you really want it to sit unused? No - you want any app or operating system task to take advantage of it so things can move faster working from RAM.
The key here - is that the OS and/or app has good memory management so that while something might have consumed a lot of RAM to be efficient, it has the sense to reduce what it is using to free it up for a task that *needs* that RAM now and has higher priority.
This is true a degree, but I have frequently run into the “out of memory, please close applications” error frequently in the past so there is, of course, a limit.
you must be using a PC as I have gotten that on PC's, but I have not on any Macs. I believe a little system clean up and maintenance will fix that for you.
nope, im on mac. And no system clean up won’t fix it, its just the apps are consuming more ram than is available. I've run into this on fresh Macs with hardly anything installed. (I haven't run into it at all on my one with 96gb but on my 16gb M1 MBP I was issued for work it commonly shows when I have a dev environment running)
Just for a matter of interest what are you doing with 96GB Ram?
I work with high-resolution complex composites in Photoshop. I find the ram tends to spike highest when I'm rendering some particle effect or working with some sort of high-fidelity blur effect while in a document with a ton of layers.
3D modeling can also use a lot of RAM. I have a few that take just under 50GB(!).
Since I bought my M1 I think that 4.5K for a laptop is kind of OK. 5K is too much. Probably till my next upgrade to M6...
If you go on the Macbook Pro configurator, you'll see Apple charges $1000 for 4TB of PCI-E 4.0 storage. Any day of the week you can find a 4TB NVME PCI-E 4.0 stick of SSD for about $120-$200.
There's nothing special about the NAND flash Apple uses. They're just jacking up prices 5-10X because what are you going to do about it? Use Windows? For the average Apple consumer that's highly unlikely.
Apple is using TLC NAND while charging what should be 2 bit MLC pricing. Though currently outside of enterprise drives, everyone has largely moved to TLC, though many modern NVMe SSDs in the $200 range, offer better performance than what Apple is offering. For example, currently the WD Black SN850X offers fasteread and write speeds, and lower latency. For 4TB of NVMe storage that will saturate a PCIe 4.0 X4 connection, you will be spending about $350 (which will be faster than Apple's offerings).
And while not really available on laptops yet, you can go for PCIe 5.0 NVMe SSDs that will do around 12GB/s, though they are difficult to cool.
Jon The Baptist thats really not true.. if you compare drives with similar transfer speeds.. or even systems from dell, or hp wit similiar drives.. the pricing isnt that dramatically different.
If you would wait a few months prices will most likely come down. It has happened with the M1 and M2 MacBook Pros as well.
When they launch a new chip family there are always production issues, the yield is low, some chips come out better than others. In many of them they have to disable a number cores as they are just broken. This is the real reason for the large variety of skews and prices.
As production stabilizes and the yield increases, prices will go down as they will not have to put such a premium on the good chips.
It is a clever Apple pricing strategy. The RAM price is ridiculous. What’s good about Apple is all new machines tend to be basically good (if expensive). Windows machine are so varied it’s very easy for the uninitiated to buy a bad machine. Intel chip naming is nuts all sorts of generations are being sold in new laptops. The Apple operating system I find awful (especially compared to how intuitive their phones are). I’m a dual Apple/Windows user. Windows isn’t great but it’s much better than Mac OS. It would be good for consumers and the environment if EU/US governments would force Apple to provide user replaceable RAM and batteries. Ten years ago batteries were easily accessible and replaceable. Now they tend to be in the most inaccessible places .
hmmm, I see why some people shoot the messenger... ouch!
For windows, it is quite efficient (with the exception of windows 11). In terms of windows 10, a standard install with all major drivers installed, windows 10 will use about 400MB for the kernel and other core aspects that load before the welcome screen, around 150MB for the GUI and it's dependancies, and around another 200MB or so for the services and dependencies. The rest after that is cached/prefetched data to ensure that things stay responsive. After that windows will cache anything that loads up until the RAM is filled, though it will only count the usage percentage based on active use. From standpoint of memory usage, windows 10 is more efficient, though windows 11 still behaves like a beta operating system.
Outside of a few technical aspects, OS choice comes down to user preference.
I think the 16" M1 Pro is still the best value option for creative pros. It might be an unpopular opinion but I think the unbinned M3 Max offers pretty good value as well.
Completely agree, the overall performance differences are small in real world between the M2 and M3 version now. For high end video then perhaps the upgrades are worth it.
I'm surprised that you're saying you do not need a ton of graphics performance as I feel that graphics performance is getting more and more important in photo editing (and presumably in video editing too).
But OTOH, perhaps the graphics performance of the higher end chips is overkill for speeding up most editing tasks beyond what the Pro model already gives. That I wouldn't know.
buy the right tool for the job.. or buy the product thats within your budget..
but it is not an apples to apples comparison (no pun intended) since apple move its own silicon.
apples ram (and everything else) is all on one chip.. meaning it is dramatically faster than any PC for that reason alone. you dont need as much ram as a comparable PC system to do the same task. (yes more ram is always better)
the storage for mac is on board.. its not equivalent to buying an NVE drive. trransferrates are much much higher.
buying a mac, you're buying a system that was completely designed by one company (hardware and software).. it makes a difference. their hardware is incredibly reliable.. and stable. support is very very good too. for a business professional thats worth a little more money.
if you dont need that.. or dont have the budget for it.. you have options. (windows, linux, etc)
much like if yu dont want to pay for lightroom/photoshop/capture one you have options too.
As a photographer working with multiple apps to producing large, high-res, complex, multi-layered composites and photobooks, I would love to understand - with a like-for-like comparisons - how you substantiate the view that you don't need as much of the faster Apple RAM as a comparable PC, or presumably(?) an Intel MBP.
Apple has been trying to convince me that the replacement they're offering for my faulty Intel MBP will be faster than what I'm used to due to the integrated RAM etc. - despite having only half the RAM and no separate GPU and despite my complex workflow and need for multiple graphics heavy applications to be open simultaneously and the maxed-out spec with separate graphics card that handled it previously.
I've been scouring the net for proper analysis with hard statistics (as opposed to the plethora of 'reviews' that mainly quote available specs, Apple's marketing hype and describe these machines as 'blazingly fast' without tangible points of comparison relevant to photographers.
They're also trying to convince me that if/when RAM insufficient and m/c starts swapping to SSD, it won't matter because writing to SSD is allegedly as fast as using RAM. I'm no Tech, but I'm unconvinced - not least because if this were true it would surely eliminate need for RAM altogether(?), but also b/c both Genius Bar tech and Apple Senior Support basically agreed I'd need the same amount of RAM as before to maintain sanity.
I found a few YouTube tech geeks who independently review h/w they purchase (as opposed to h/w they're given to review) calling bullshit on the whole theory, insisting Apple's RAM isn't magic and the same amount of RAM is required on M3 as Intel Mac - if not more, because it's now shared with graphics as well as O/S - but if you can enlighten me, I'd really appreciate it.
Apple includes all the RAM on the M3 chipset, right? So why would it cost more to make?
They basically solder the DRAM packages very close to the SOC. It is a similar implementation to how VRAM is added to a video card. Direct soldering of the BGA packages as close as possible to the SOC, manes it easier to have faster memory that can run stably at lower voltages. From a cost standpoint, it is cheaper to go with a BGA DRAM package than having a separate DIMM module., though it also makes it nearly impossible to upgrade the RAM.
Yeah, apple is Apple and I will spend 35 times more just because is Magical and Genius. Also, have you seen how mindblowing and headturning it is? Evil grifter company that is.
I use them too, but their pricing for added RAM is designed strictly for maximizing revenue. I know, shocking. But really, it's a rip. I don't believe changing the RAM on that chipset costs them a single dime more. Hell, it's probably already there, but limited somehow.
putting everything on one chip doesnt make it cheaper.. it makes it faster and more efficient.. all on one chip is far more complicated.. and if any of those parts doesnt work during manufacturing.. the whole chip is bad and cant be used. (more expensive)
The difference these days with computers is fairly easy to determine using benchmarks but not as easy to see using real world applications. I just wonder what the real world difference is between an m2 and m3 machine. 8GB of memory in these M3 machines is a joke. For those that browse the web and use standard office like apps I fail to see what the point is upgrading if indeed people do that. A MacBook from a decade or more ago is more than easily capable of doing what the vast majority of users want to do. While editing multiple streams of log footage is certainly taxing for any machine just how many users need that level of performance? The prices of these higher spec M3s is eye watering indeed. Apple are really milking it or should I say us, or me! for all it’s worth having just invested in a new Mac Studio.
Just a general word about Apple. I’ve been a user since 1985, Mac plus Se etc. Once you become used to the Apple environment you tend to stick. Why? The vast majority of the time everything just works and keeps working, and working. Reliability is key. Now with phones and watches. I’ve just acquired 15Pro Max and Ultra 2. Apple have a long time locked in user base to sell to, old guys with too much cash and not enough sense, as well as attracting new users. In all honesty I preferred the simplicity of System7 and earlier. The now bloated and capable Sonoma is for me way beyond overkill. I don’t need or want many of its features. I just wonder what difference a slimed down system would make? Though you do have to take your hat off to the engineers and designers at Apple. I unboxed my new phone the other day, old one an Xr. Switched it on brought the old one close as requested by the new phone and that was that. The process was seamless and easy. Everything is under the hood and just works, and that is Apple in a nutshell. Thats why people keep coming back. The downside is it’s not cheap, but look at any Apple hardware and it reeks quality. You do get what you pay for but there is a premium to pay which many like myself are prepared to shell out for. Though when it came to a new monitor for my MacStudio I had to give Apple a body swerve, I am crazy but not that crazy, and instead went for the BenQ SW321C.
At the premium market segment it is targeting, it should start at 32GB. 8GB is unusable, and 16GB is only really useful for basic content consumption. Many M1 and M2 macbooks ended up with insanely high write amounts on their SSD due to the constsnt use of virtual memory/ swap for most tasks, and those issues will only be worse now with applications becoming even more memory hungry.
At even $1600, 32GB should be the base config, with a $100-$150 option to upgrade to 64GB.
have you experienced this virtual memory swap yourself and is it the pain that I assume it would be asking because Apple are trying to convince me it's seamless and undetectable.
With any system, once you start relying on virtual memory, performance will slow to a crawl. The only time when the process is not very noticeable, is if an application does not need to actively use more than the dedicated RAM. Keep in mind that on the Apple M1, M2, and M3, the RAM is also shared with the GPU, thus if they lost 8GB, then you can expect at least 2GB of that to be allocated to the GPU when using a program like a video editor.
The type of swapping that is relatively seamless, is when the OS moves the prefetch and soem background applications onto the SSD to free up more DRAM for the application. But if you have a situation where the OS can only make 5 to 6GB of RAM available to the app, but the app needs 10GB, then it will constantly be using the SSD, which is only a tiny fraction of the performance.
Programs like photoshop are more forgiving to limited RAM, since it can work within a shared 8GB without constantly swapping unless you use lots of layers and smart objects, but programs like DaVinci Resolve are not forgiving for such limited DRAM scenarios.
Thanks for responding. I'm a stills photographer and I don't use video, but I do run multiple Adobe apps simultaneously (each requiring 15GB according to Adobe), plus the usual background stuff (multiple tabs, mail, messaging apps, sometimes MS Word / PPT too). I import large quantities of high res RAW photo files into Lightroom (eg 700 x 50MB), which are then edited between Lightroom, Photoshop and Topaz. All apps now utilise AI in processing which can be slow. I tend to work on multiple files at the same time in Photoshop (say 2-8), many of which are large and multi-layered, up to 1-2GB each. I occasionally do billboard size high res images. My worry is that although my old Intel i9 MBP with 64GB and seperate 8GB GPU (so effectively 72 GB RAM) coped well, the new M3 Max with only 48GB might start swapping when I've got 3 x 15GB apps open, a huge LR catalog, no seperate graphics card and the OS sharing that RAM too. I'm no tech and Apple insists I'm wrong... but am I?
In such a use case, if each app is being actively used for larger files at the same time, then swapping will take place. I currently make due with 64GB of RAM (moving to 128GB would require me to go with higher latency RAM modules, which will negativity impact gaming performance), and will use DaVinci Resolve, photoshop, and sometimes Adobe premiere CC. Though it is a struggle, especially with my videocard only having 16GB of VRAM which can get limited with complex node trees, especially in their fusion tab. While a card like a Radeon RX 7900 XT would have been better for this workload, I needed a few Nvidia specific features for other applications.
Anyway, there is no special solution that will allow a system to get away with having less RAM. While many applications will try to work within the limits of available RAM, it will always negatively impact performance. Compared to windows 11, Apple does better with memory management, in that they are more used to developing for limited hardware, while the design culture surrounding windows has been to throw more hardware at the issue. With that in mind, the difference isn't massive, instead it is more of impacts at the borderline hardware. For example, if you plan to use a laptop for mainly media consumption, 8GB on a MacBook can work fine, but on windows 11, it will struggle since the OS wants to use nearly 4. 3GB just on itself at startup. Outside of that, RAM caching and swapping will behave similarly between windows and Mac OS.
Thanks Naruto!