"AI-Focused" means "32 core neural engine"
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.
Macs to Get AI-Focused M4 Chips Starting in Late 2024
- Thread starter MacRumors
- Start date
- Sort by reaction score
Is this an serious question? When speed was a bad thing? Sell sell if you need, nobody is forcing you.Are you talking about M3 vs M4 alone or also the entire M family like M4 pro Max ultra
For me, that i work with medium maya projects , every 20% improvements means less time to develop...less time mean 2 things to chose from, more free time, or more projects
Well, let’s look at the last several redesigns…
First aluminum PowerBook G4 design: lasted from October 2003 to October 2008/January 2009
First unibody design: introduced October 2008, replaced in June 2012, not discontinued until 2016.
First RMBP design: introduced in June/October 2012, replace in October 2016
Touch bar design: introduced in October 2016, replaced in October 2021, discontinued in October 2023.
The current design, going off of past patterns, should still be around at least another year or two.
A redesigned OLED (possibly Touchscreen) MacBook Pro is expected in 2026-27.
yep, they just stated the m4 studio was coming out real soon... that prediction didnt have much of a shelf life.
I’m already past due to to update to my desktop. Was gonna finally get an m3 studio this summer but now i dunno lol. M3 is not turning out to be the revolution that was expected
HobeSoundDarryl
macrumors G5
...and probably Studio and Pro.
I need a big fat incentive to upgrade from M1. Is some sort of on-device GPT-Siri that meaningful?
I sure hope what Apple is cooking up isn't just a copy of what everybody else is doing, and that we're supposed to be impressed because its happening on-device and respects privacy.
Sure, that's great.
But I need it to do something different and better than the others, not just that the "behind-the-scenes" is technically more impressive.
It has to be immediately and obviously impressive.
I sure hope what Apple is cooking up isn't just a copy of what everybody else is doing, and that we're supposed to be impressed because its happening on-device and respects privacy.
Sure, that's great.
But I need it to do something different and better than the others, not just that the "behind-the-scenes" is technically more impressive.
It has to be immediately and obviously impressive.
How about we get removing-the-notch focused chips?
It means running local LLMs models that can help with daily tasks (like MS Copilot), but maybe, just maybe, Apple could make that better. MS Copilot is like Clippy with steroids, not really that useful.
HobeSoundDarryl
macrumors G5
They already posted an 11-minute video of this a few days ago... headline being a potential flip in the order of rollouts, beginning with "ultra" and getting to "base" instead of "ultra" rolling out last as it has been. Here it is...
The key speculation is first M4 will be (N3E-based) M4 Ultra in Studio and Pro this summer, followed by MAX & PRO in MBpro this fall, followed by base M4 in the Macs that lean on it late 2024/early 2025. Why? One of the bigger keys is that there is more profit in the higher-end chips and rolling them out last motivated people to WAIT for the next generation which was always impending soon after Ultra. Next gen MAX has generally been close to prior gen ULTRA in key tests, so there was even some consumer sense in waiting and choosing MAX over ULTRA. Most simply: the window of time for ULTRA to be "most powerful Mac" has been relatively tight. Flipping the launch schedule makes it king for the entire year or so until the next ULTRA is released.
By flipping it to Ultra first-base last, those wanting "latest & greatest" will pay maximum money (and profit) for it and then it will step down to the less profitable versions of the same chip. Base will be rolling out close to the new next generation, so those most driven by "latest & greatest" would feel some pull to wait and then pay up for Ultra... or then MAX & PRO.
Whether there is merit to this speculation or not, when I see any rationalization that revolves around "more profit" for modern Apple Inc... I at least consider it more seriously than just dismissing it as a wild guess.
Last edited:
Since Google is publicly stating that their future smartphone will require 24GB of RAM to run Artificial intelligence stuff — one can only guess how much RAM a regular computer will need to enter the ai realm
Apple’s on device chips use snippets of machine learning code to execute certain tasks: voice recognition, image recognition (try searching your photo library, it’s on your device), image creation (all your photo’s go through that ML code), etc etc. This is different (and in addition to) cloud based services such as ChatGPT. Advantages for on-device code: speed, no data over the internet and privacy. Disadvantage: limited memory for code.
Apple has been doing this for years, nothing new.
Apple is likely looking for a sweet-spot between on-device and cloud based computing. This will differentiate them from Google and Microsoft which mostly offer cloud-based computing.
8GB, take it or leave it!
This is my big question. I wonder if the m4 pro/max is for the OLED MacBook Pro
I still think mini, mini Pro, and Studio will see M3 versions at WWDC. Probably also Mac Pro. It's about the same copy paste spec bump that Apple does for every generation and shouldn't cost them much.
But if mini, mini Pro, Mac Studio and Mac Pro skip M3 altogether then we'll see M4 versions of those no earlier than WWDC 2025.
Post every annual iPhone event, MBP and MBA are the Macs first in line to get the new chip generation so it would only be those two moving to M4 in late 2024 or early 2025.
AI....AI...if AI wasn't as hyped up in the industry as it is, I'm sure the normal cadence would've led to these models getting the M3 chips...
Still using my 2019 iMac with an Intel i9 processor and 64gb of ram. Would like to upgrade if Apple provides a machine with the desired specifications.
1. All in one design.
2. A screen size of approximately 30 inches. (Shrink the bezels on the old 27 inch iMac).
3. Processor equivalent to the then current Macbook Pro.
4. A good quality FaceTime camera.
5. SD card reader, Audio port, USB A and Thunderbolt ports.
6. A reasonable price. The studio is plain and simple overpriced in comparison to the old 27 inch iMac's.
1. All in one design.
2. A screen size of approximately 30 inches. (Shrink the bezels on the old 27 inch iMac).
3. Processor equivalent to the then current Macbook Pro.
4. A good quality FaceTime camera.
5. SD card reader, Audio port, USB A and Thunderbolt ports.
6. A reasonable price. The studio is plain and simple overpriced in comparison to the old 27 inch iMac's.
I wouldn't obsess over performance gains. The goal is more optimization of the user experience. Putting that into hardware, rather than software will definitely speed the response times of any user interaction with his computer. For example, instead of giving Siri one command at a time, imagine leaving your home and having a shortcut called "run 'away" sequence." Siri then lowers the thermostat, sets away scheduling for house lights, turns the TV on/off at random, activates the home alarm, checks to be sure all doors are locked, halts your mail or newspaper delivery, pays bills when they come due, adds an automated "out of the office" reply to any emails, goes online and gets your boarding passes if flying, or starts your car , turns on the A/C or heat, and unlocks the car doors." And that's just a tiny sample of the potential.
sad that apple had to succumb to the marketing slop of AI