Yesterday Tim Cook said: “We continue to feel very bullish about our opportunity in generative AI and we’re making significant investments,”
Currently all the big players are offering tools that run in the cloud. It will be interesting to hear what Apple has planned, especially if they plan to stay on-device.
Yeah, I’m very curious about this too. Cloud-based computing is fast until your network is not. I like local computing for that reason.
Local is also much, much more private.
I agree.
I use an M1 Air and M1 mini, both are still brilliant. I use the Air for work, presentations, (large) spreadsheets, writing and what have you. I sometimes use it for photo editing.
I will grant you I don’t do video editing, 3D modelling or programming, but I have never once felt the need for more power, never seen a slow down, never had a reliability problem. I did borrow an M3 air for a short period (same base spec) and while the design was stunning I really didn’t notice the upgraded power for the work I do.
These M class machines are amazing.
I agree the M1 is such a massive step (it’s where Apple after many years finally released the laptop I want) from Intel a bargain can be found right now. You’re on Intel so do not need an M3 Pro right now.
Been trying to convince myself to mover from an M1 to M3 Pro - however there is absolutely no sensible reason.
And to add to this, the M2 doesn’t feel significantly faster than the M1.
Intel to M1 = 
M1 to M2 = 
M2 to M3 = (We shall see 
)
If Apple on Tuesday would announce an iPad Pro that also runs macOS, the right time might be now…
The only reason I would want a newer machine than my M1 MacBook Air is that when processing photos with Lightroom Classic, Adobe Bridge, FileLoupe, and PhotoMechanic, it is easy to go past the 16 GB of memory that I am limited to with an M1 Mac. (I would get 32 GB of memory next time.)
The funny thing is, the M1 Mac takes it in stride with barely a hiccup, kicks some stuff out of memory onto my internal SSD as operating system swap space, and continues almost as if nothing has happened So even then, I really don’t need more memory! 
I believe the MBA maxes at 24 GB, so currently you’d have to get a MBP if 32 is important to you, though of course that may change by the time you decide to actually replace your current laptop.
My M1 MBA only went up to 16 GB. If I want 32GB in my next Mac, you are correct that I would have to get something other than a MacBook Air. (P.S. I knew that )
MacBook Air (M1, 2020) - Technical Specifications - Apple Support
16GB unified memory
512GB, 1TB, or 2TB SSD
MacBook Air (M2, 2022) - Technical Specifications - Apple Support
16GB or 24GB unified memory
512GB, 1TB, or 2TB SSD
MacBook Air (13-inch, M3, 2024) - Technical Specifications - Apple Support
16GB or 24GB unified memory
512GB, 1TB, or 2TB SSD
Yes, that was true of the 13” M1 MBP, too. Someone I knew bought one, and I recommended they max it out at 16 GB.
I have a 13-inch M1 MacBook Air. Max memory was, and still is, 16 GB. Bought it long before an M2 or M3 MBA was available with 24 GB.
I had considered getting a new MacBook Air once Apple upped the max memory to 24 GB, but 24 GB memory didn’t feel like enough headroom to me. A MacBook Air is plenty of machine for me from a processor and SSD point of view, but I wouldn’t want to be bumping up against a memory ceiling again any time soon. So, it seems, I will have to “Go Pro” with my next Mac.
Just a bit of historic reference, when IBM invented and introduced Virtual Memory in the 1960’s, it was believed that it would save money as customers would need less real memory on their computers. So applications got larger, and the programs started to swap, and performance dropped precipitously. Customers were forced to buy more memory than ever!
Virtual Memory was viewed to be the best ploy ever to sell more memory!
So see today’s 8GB is as good as 16GB scam.
Except of course, these days you can’t install extra memory after your initial purchase.
FWIW, I upgraded from an M1 Max to an M3 Max (16" MacBook Pro for both) and I found the M3 Max to be noticeably snappier for a lot of basic tasks. Not as big as the Intel->M1 jump obviously but still well worth it for me IMO.
Great and well-reasoned discussion. As someone who bought the first M1 MacBook Air, with 16GB of RAM and 1TB storage, it still does everything I need it to do, save train transformers, though it is amazing to watch the Activity Monitor report that the system is using 750% of the CPU.
Unless you do computational work or edit large media files (images, video, and possibly audio), your best bet is to focus on RAM and storage and not worry about which generation M is in the machine.
And, FTR, I agree with the comments that observe that, first, we’ve had ML baked into the OS for a while and I’m guessing the “neural engine” is to make the kind of computation I noted above easier to run on Mac hardware, and that ML is generally ML until it moves into the public sphere and then it becomes AI.
Similar, using a mid-2014 Intel MacBook Pro for non-work use. I’ve learned to work around the 1-2hr battery life and 128Gb storage buy using it on a desktop with power and external SSDs.
My buying advice for a personal purchase is to consider what is needed against what’s wanted. Unless device dies or prevents an essential use case this is generally a want rather than need.
My experience is that the hardware is not the limitation. Instead my principle issue is the number of apps which demand the latest OS. I understand the support challenge for multiple older OS. However the lack of these apps for macOS (in my case) actually pushes me to alternative Mac workflows / services, or to only iOS or web.
For a while I have wanted a new MacBook to address these shortcomings. At this point I’m weighing up a new MacBook this year, but may defer it a little longer if the M4 is just around the corner and given that I don’t need to replace it.
9to5 Mac says Apple has a much better manufacturing process with higher yields for the M4 than for the M3, and that’s why they’re moving to the M4 so quickly. So we may see an M4 MacBook Air sooner than expected, though of course that’s just conjecture.
The reasons behind the sooner-than-expected launch of the M4 are probably correct and are not much of a surprise given what we knew from the leaks about the 3nm tech manufacturing process, but it is sometimes neglected that the neural engine in M4 received a significant upgrade for a laptop chip (this part being vital here). The list below is from a reddit thread. While Apple has touted this at the keynote, this signifies a change in strategy (possibly, and I’d even say probably influenced by the AI craze).
Note how Apple has so far always put a more performant chip – in terms of ML/AI/neural engine/whatever they wish to call it – in iPhones to power their ML features, probably assuming that laptops have enough power as is (from their CPU & GPU) to do these tasks when needed, and how M3 over M2 was just a minor upgrade here.
A17 Pro, for example, outperforms the M3 – which is the same chip generation – by a factor of two (!), and the A16 outperforms even the M2 Max. This trend seems to change with the M4 with more than double the neural engine performance from a previous laptop chip (now just over six months old). I’m sure the A18 will again receive a major upgrade here, but a laptop chip has never received such an upgrade of its neural engine.
- A14 Bionic (iPad 10): 11 Trillion operations per second (OPS)
- A15 Bionic (iPhone SE/13/14/14 Plus, iPad mini 6): 15.8 Trillion OPS
- M2, M2 Pro, M2 Max (iPad Air, Vision Pro, MacBook Air, Mac mini, Mac Studio): 15.8 Trillion OPS
- A16 Bionic (iPhone 15/15 Plus): 17 Trillion OPS
- M3, M3 Pro, M3 Max (iMac, MacBook Air, MacBook Pro): 18 Trillion OPS
- M2 Ultra (Mac Studio, Mac Pro): 31.6 Trillion OPS
- A17 Pro (iPhone 15 Pro/Pro Max): 35 Trillion OPS
- M4 (iPad Pro 2024): 38 Trillion OPS
This could pretty much influence which devices support whatever ML/AI features Apple decides to introduce with iOS 18 and macOS 15. Some advanced ones could likely be reserved for the A17 Pro and M4.
I understand that. And I’ve debated for a while if i could downgrade to an Air for my needs. My only option there may be to purchase an Air with a view to a potential return and replacement with a Pro.