It isn't the most powerful computer in the world but I never ran into any problems... so it's probably an OK compromise for most people, especially in the world where RAM is scarce because of AI datacenter buildouts.
Shortly after I had a TiBook (PowerBook G4) that was _only_ 1-inch thick! Compared to 1.75” Dells my coworkers had, it seemed like the future. DVD drive, modem, Ethernet, full sized DVI port, FireWire, WiFi, Bluetooth, optical audio in and out, gigantic display with a bezel that was unrivaled for years, even among Macs. What a beast!
(I know you meant 2020, but it’s fun to think about the air in 2000).
https://top500.org/lists/top500/list/2000/06/
https://www.notebookcheck.net/Apple-M1-GPU-Benchmarks-and-Sp...
Only true if all you're running is matmul (supercomputer has general purpose CPUs so more flexible than M1 GPU) - also those flops are probably FP64 in supercomputer ratings and FP32 in M1.
As a smart man I knew used to say, supercomputers are about I/O not raw compute. Those have terabytes of RAM not 8GB.
Throughput-wise, the supercomputer is competitive because it has a lot of local RAM connected to lots of independent nodes, which, in aggregate, is comparable to modern laptop's RAM throughput (still much more than disk) with a caveat, that you can only leverage the supercomputer bandwidth if your workload is embarrassingly parallel running on all nodes[1]. Latency-wise, old RAM still beats NVMe by two or three orders of magnitude.
[1]: there's another advantage that supercomputer has which is lots more of local SRAM caches. If the workload is parallel and can benefit from cache locality, it blows away the modern microprocessor.
It was also used by the PS2 for local multiplayer between multiple consoles. Although Sony eventually removed that port.
All our intel MacBooks now run Linux just fine. The oldest is 2012, with 4Gb but most are 8 or 16Gb.
I would always recommend more RAM first over a faster processor; back when I would build desktop machines for Windows, I would use the second best CPU and put the savings into RAM.
OTOH, for my development Mac, I have 64GB of RAM. (Though 32GB would probably be fine.)
OTOH, for my development Mac, I have 64GB of RAM.
(Though 32GB would probably be fine.)
Definitely brings our engineers' 24GB MBPs to their knees primarily b/c of the RAM chewed up by those multiple Docker instances.
Will 32GB also start looking paltry soon? It's hard to say. I want to say the realistic upper limit is 3-4 simultaneous worktrees for a given developer (at this point the developer becomes the bottleneck again?) but it's a wild guess that may be hilariously low.
(Just posting this to show that you have to be very specific when talking about these kind of things. Yeah maybe you need 32GB because you run some large deployment 3 times. Others mayb be totally fine with less if they just develop a basic Python web app. Who knows. The devil is in the details. Omitting them makes the discussion ambiguous and just difficult.)
Students in non STEM areas will not usually need more than writing and reading tools.
Mac devices have been able to get away with less RAM (and higher priced upgrades) for well over a decade. During the Intel era, they were the first ones to adopt SSDs as the default option while everyone else still installed spinning rust. That alone provides for way faster swap storage to conceal a relative "lack" of RAM.
And when they went for their own fully integrated stacks of soldered RAM and SSD? Then everything went off the rails - close proximity and no sockets means very low latency for both RAM and persistent storage on one side and on the other side it also allows for much higher bandwidth because of much cleaner signals - remember, even at "measly" hundreds of megahertz you're already in the territory requiring precise PCB design.
On top of that, macOS's scheduler seems to be much, much more efficient and outright better in constrained RAM (and CPU) settings to provide the feeling of "the system is still responding" than either Windows or Linux. The only setting where macOS goes into molasses is when you not just run out of RAM but of free disk space as well.
I’m not saying you can’t get away with 8GB of RAM. You can, but I won’t recommend a Mac with only 8GB of RAM to anybody for a few reasons: 1) even normal users just using the web will find RAM to be the bottleneck and that will degrade their user experience over time. 2) they’ll spend $600 and even if RAM isn’t as much of a bottleneck for them today, with modern web developers and modern web browsers, it will be much sooner rather than later. And everything is a web app now.
For $600 in 2026, your computer shouldn’t be a bad experience in any way nor should it last less than 7 years and still be a kickass experience. Ideally it should last longer. The Neo is great for what it is, but the RAM is the deal killer for me.
(Been trying to get them to switch to Safari, but they prefer the Firefox name. I don't think there's anything wrong with Firefox other than it being less native.)
A 20-tab Chrome requires 50% more RAM than Safari.
What actually kicked off my browser exploration on the personal was dissatisfaction with Safari’s performance, and 20 tabs or less was enough to make it drag at the time even with disciplined use. I don’t think it had any significant advantages over a Chromium-based browser that particular year except probably battery life but battery life has not been an issue for me these entire 5 years. RAM and swap are something I do end up monitoring more each year (and I’m not in Tahoe yet for either of them), but I’m planning to drive these into the ground before replacing them.
Reminds me of a childhood friend of mine who always said "it looks very 3D" when he meant "the graphics are good". Pissed me off back then, and apparently still does.
It's actually enabling my worst tab-hoarding tendencies. In the Intel days I'd pay a performance price at some point and have to tend to my tabs, but now they just keep propagating....
It's basically a Laptop engineered in the iPhone/iPad space of the company, it's only natural for Apple to target a shorter lifecycle.
8GB RAM is maybe the best way to achieve that, many of the MacBook Neo buyers of today will be very compelled to upgrade to a newer (or higher-tier) model in ~3 years from now...
If the Neo would have 16GB of RAM today, it would be harder to justify an upgrade in 3 years from now, when the common entry-tier for laptops is likely still at 16GB...
> “For $600 in 2026, your computer shouldn’t be a bad experience in any way”
In the article, Gruber normally uses a 64GB Mac, expected the 8GB RAM to be a problem and was surprised to find that it wasn’t, and judged the Neo as not being a bad experience in any way.
8GB can be limiting on an iPad Pro, which runs a generally more memory efficient branch of Apple’s XNU-based system software and it’s not difficult to get it into a state where it is constantly paging out an app you had in front of you two minute ago if anything you’re doing involves the web at all. A Mac will just swap at that point, but swapping is also slow.
Five years from now, I have no doubt that the processor will still be fine for most uses, but I doubt that 8GB will be. Especially given that some of the most common memory hogs aren't under Apple's control (cough Chrome cough).
If it's not even useful after such a short time, then I question whether it was really fit for the intended purpose even when it was new.
Software shouldn't be written this way. I shouldn't have to disable mds-store because it likes to take up 2-3 cores at full throttle when I'm on 10% remaining battery. But it is, and 32GB isn't enough for me to even have a basic computing experience anymore, it seems.
Yes and no. I had a M1 MacBook Air for several years, with 8 GB. It's fine if your needs are relatively simple (ie: just a browser, with not too many tabs, and a few other simple apps). But try to run too many apps and it would tend to hit a wall and get very slow.
One thing that did seem to help a lot was to keep the SSD relatively empty: the SSD seems to get slower once it has < 30% or so free space remaining, which would slow the whole system down because memory swapping takes longer)
Last year I upgraded to an M4 Air and got 24GB, which makes a world of difference. But I gave the M1 to my niece and she seems very happy with it!
You and I disagree on this part so strenuously I don’t foresee a middle ground. Swap still absolutely sucks no matter how fast the SSD is, and the SSDs or probably the SSD controller are much slower than what’s in other Apple Silicon Macs.
People always forget that Apple does realtime compression on data that's in RAM allowing more things to fit in RAM; it also effectively increases the bandwidth of the SSD.
I have a 16GB M1 Pro machine from 2021 with 200 GB/s memory bandwidth; I can't tell when it's hitting swap, even with tons of browser tabs open, 3 or 4 terminal sessions, and several apps running. I often run two browsers with dozens of tabs open and there's no noticeable lag.
YMMV.
On an my old Intel Mac, it was pretty obvious.
Although, I guess Windows 3.1 and 95 users enjoyed it first thanks to this extremely high quality third-party implementation!
A combination of Apple's Unified Memory Architecture (UMA) and hardware-accelerated instructions (SIMD/NEON) makes RAM compression on Macs very efficient. Because the storage controller is integrated into the SoC, the bandwidth is high enough that the transition between "Compressed RAM" and "Swap" is very smooth.
And because the CPU and GPU share the same memory, there are no wasted cycles moving data between VRAM and System RAM.
Apple uses WKDM (Wilson-Kaplan Direct Mapping), a specialized, high-speed compression algorithm designed specifically for in-memory data. WKDM is "architecturally aware"—it was built to compress the specific types of data structures found in a computer's RAM, such as pointers, integers, and memory addresses. WKDM treats RAM like a collection of 64-bit integers and pointers; and it's designed to fit entirely in L1/L2 cache [1]. This shipped in MacOS 10.9 Mavericks in 2013.
Windows/Linux treat RAM like a stream of bytes (similar to how you’d compress a .zip file) so it’s not as efficient. The vast majority of Windows and Linux machines don't have unified memory or storage controllers connected to their processors.
Because of this, Apple can often compress a page of memory using fewer CPU cycles than Windows or Linux, which is why M-series Macs can be so aggressive with compression without you ever noticing a "hitch" in the UI.
The fallback algorithm is their LZFSE algorithm, which is like "Zlib-level compression with 2x-3x the speed and efficiency". LZFSE achieves a nearly identical compression ratio but uses Finite State Entropy (FSE) coding, which allows it to decompress data significantly faster while using much less battery power.
LZFSE is optimized for the ARM NEON instruction set to minimize "wake time" for the CPU, making it arguably the more "green" choice for mobile devices [2].
It's safe to say that neither Windows nor Linux has the combination of hardware and software optimizations that Apple has when it comes to RAM compression.
[1]: Compressed Memory compresses the least recently used data residing in memory using the WKDM algorithm, which not only frees up memory but also reduces the amount of swapping going in the background. Not only is this faster than swapping to disk (even to SSDs), but Apple also claims it saves power -- essentially, that compressing data in memory uses less power than writing data to disk without compressing it. -- https://www.osnews.com/story/27121/os-x-109-mavericks/#:~:te...
[2]: https://lyncd.com/2015/09/lossless-compression-innovation/
> Windows/Linux treat RAM like a stream of bytes (similar to how you’d compress a .zip file) so it’s not as efficient.
That doesn't really follow. There are faster and slower compression algorithms no matter what, and 64-bit integers are kind of a waste of memory much of the time.
Also, unified memory has tradeoffs. The GPU improvements are real but it mostly means more pressured on memory, not less.
That's not in the same universe as hardware compression on a 6-core, 64-bit ARM processor with cores that can run at 4GHz.
OTOH, if you are using multiple RAM-heavy apps that aren't actively hammering that RAM (e.g. an instance of Photoshop that is using 10GB but is just idling or whatever) then MacOS and their stupid fast SSDs handle that pretty seamlessly.
Most use cases are probably somewhere in the middle.
The best comparator here is likely the iPhone 16 Pros, released in late 2024. These were the flagship iPhones until late 2025. They are only one generation old. They have the same CPU and the same 8GB of RAM. I have never heard anyone complain that they suffer performance-wise from having too little RAM.
Many of the apps non-devs use will likely be universal binaries, or adapted from iOS versions. Chrome, Safari, Slack, Calendar, Gmail, Zoom, Claude, Contacts, Notes, Maps, Music, Pages, Numbers, etc. These are apps that run concurrently with no issues on the iPhone Pro 16. I'm not sure why people expect those same apps would cause issues on materially the same hardware because its package includes a hardware keyboard.
(The most RAM you could purchase in an iPhone until late 2024 was 6GB. iPhone 11 had 4 GB of RAM. I have not at any point since approximately iPhone 6 heard anyone complain about the speed of an iPhone Pro for "normal" consumer/not professional media stuff. iPhone 6s was released in late 2015 and had 2GB of RAM.)
Yes, MacOS is a different OS than iOS. But the very same company who built the Neo also make MacOS. They are known to adapt the OS to the hardware they are shipping. I'm willing to bet the experience for the non-dev is similar to the experience of using an iPhone 16 Pro in 2026.
So, you can't really compare. On iOS you can have 3GB of RAM and it wouldn't be a bottleneck.
Except 1) that's not entirely true (famously: music, Zoom) and 2) yes, cooperative state management. Users do not know or care that an app is not actually running if it appears that it is still running when they switch back to it. #2 obviously does not work for many dev use cases, but it would not impact my workflow if e.g. ChatGPT or Chrome were suspended when not in the foreground.
I have 8GB of RAM in my M2 iPad Pro running iOS (yes, it’s “iOS” despite what Apple’s crack marketing team might call it), and I’ve certainly started to complain. Doing anything with the web, and like one or two other apps is enough to have apps I’m switching between page out like every two or three minutes.
I had issues with swapping before, but with the latest iOS, it has become very annoying on an old iPhone with a small amount of RAM (3GB, I think).
Apple fanboys laugh at Android users for many things, but they can use their devices longer even though they might not have the fastest CPU around (8GB+ has been normalized forever in Android world).
I agree with most of the post's arguments, and most of the specs and limitations of the Neo would be okay with me, except there should be 16 GB RAM in 2026.
Apple could perhaps mitigate this somewhat by releasing a "slim" MacOS Neo version that is less bloated by pruning some features. Currently, the OS uses much of the available RAM for caching (I've seen "40%" of total OS RAM usage) to make the system faster, whereas 8 GB RAM permits only essential caching.
(Surely, the tough 8 GB RAM decision was influenced by the three factors 1. current DRAM cost and 2. limited DRAM availability considerations as of 2026, and 3. the massive Neo market size resulting from its attractive price tag, and this may get reconsidered in future editions.)
I think it’s as simple as: 8GB is what the iPhones using the A18 Pro had. It’s this thing Apple likes to do where to keep costs down, they use some iPhone part or other SoC/SiP they have laying around as close to its standard configuration as possible with minimal changes.
Their new Studio Displays for example have an A19 Pro and 128GB of NAND. For basically just the firmware. Why? Because that’s the least amount of storage Apple ships with an A19 Pro iPhone, because like the previous Studio Display from 2022 which had an A13 Bionic in there, they probably just shoved an iPhone board in there to handle the logic and I/O.
So in theory, if they update the MacBook Neo next year to an A19 Pro, it should have 12GB of RAM.
The reality is that nobody outside of HN cares about 8GB vs 16GB of RAM. You can do anything you want or need to do with an 8GB Macbook, including running a million dollar business, or working with anything creative on the highest level. If you are actually doing something which requires 16GB of RAM on a Mac, then you are doing state of the art tech stuff and should be rolling in money already and have no problem spending thousands and thousands on your computer.
Actually it's because the A18 Pro only supports 8GB of RAM. It's packaged on top of the SoC itself using TSMC's InFO-PoP.
That is ultimately what keeps saving Apple from turning into Dell. They want to offer you one model per price point. You'd be hard-pressed to find two iPads, Macs, iPhones with the exact same price. There's always a price difference with Apple, which helps immensely.
The Neo is either easy to recommend or rather easy to not recommend. It has a fixed 8GB of RAM. I think that’s too little for a modern Mac operating on the modern web. Others… disagree. Either way, it might entice some schools and school districts assuming they can volume discounts where 8GB is probably enough and it fills the spot in the Walmart part of the sales channel previously occupied by an 8GB RAM M1 MacBook Air Apple hadn’t sold itself in years.