I hope they're also working on redesigning the MacBook Pro to feature a display without a notch. I'd buy that instantly.
The only hiccups I ever have are resolving Swift result builder types…perhaps SwiftUI is just a big ploy to burn up all the extra CPU cycles.
Even on the fanless Air performance was more than enough for the moderate dev work I threw at it.
iPhones have been more powerful than they needed to be for years; it's great to see this phenomenon in Macs.
I’m still happy on my 2020 M1 normie MacBook Pro.
What did change is almost everything else. Much better speakers, more ports, more external displays, no Touch Bar and probably some other stuff I forgot. Not worth buying a new laptop over though.
Today I happily ordered the M4 Max with 64 GB with otherwise same specs for 4600 EUR!!! Yay! Will return the other device happily.
Thanks @Apple for a huge increase in specs for not even 100 EUR more :)
PS: I just needed the RAM, as my M2/16GB constantly was using swap memory of up to 16GB and was really slow for my taste. I don't do "heavy weight lifting" with it, just leave all apps and tabs open :D
PPS: I use VS Code, Emacs, Firefox (eats my RAM super fast :), and a ton of other apps, and in Firefox I use lots of web based apps (Figma, Lucid chart, Canva and some more)
That is if I don't just suddenly decide to replace the batteries (relatively cheap from the apple authorized store compared to a full upgrade).
https://www.theverge.com/2024/10/30/24283651/cyberpunk-2077-...
There are a few such comments here already!
So while frustrating to people that want to geek out on the tiny details, the real market for Apple is people not using Apple silicon and maybe the M1 users. Upgrading from M2 or M3 isn't worth it. Thus the comparisons with Intel Macs and M1s.
- most people don't care because they're not contemplating upgrading after one year
- those who care will find the benchmarks elsewhere
But that aside I agree it feels like marketing slime to tout improvements over 3 gens ago. At least include comparisons to last gen. Preferably with numbers on the chart.
I think theoretically you could run inference on Llama 3.1 405B (4 bit) on a Mac Studio which is kinda nuts.
Seems like the M4 Max has the same "specs" as the M3 Max. Same core count, gpu code count, and memory amount.
Is multitasking across a web browser and Excel slow on any CPU nowadays?
More seriously, though, it probably shows that CPU speed is not the bottleneck for most tasks nowadays.
