[thruhiker]

I have this exact build of MacBook Air and in day to day use browsing the web, doing some light coding, and connecting to Kubernetes clusters hosted in the cloud, I don't really notice. There is definitely some disk swapping occurring but you don't really notice it in small amounts due to how fast the SSDs are.

Would I recommend the 8GB of RAM to someone who wants to use it as their primary development machine or for heavy photo editing, absolutely not. Would I say it's OK for casual use, yes.

Do I think Apple is being stingy considering their incremental cost for an extra 4GB or 8GB of RAM in the base model would be tiny, yes.

[art0rz]

I bought my gf an M3 pro with 8gb ram and text input in the browser is often slow as hell (Google Docs, Facebook Messenger, etc.). Takes a moment before a character appears after typing it. I've never owned a Mac myself so I believed Apple's marketing speak about 8gb being the 16gb on their platform (not to mention that the 16gb upgrade is ridiculously expensive) but boy are they wrong. Maybe in native applications their memory magic works but definitely not in a browser. She does some light React dev and it's been fine for that though (but not mind-blowing).

[robenkleene]

This isn't enough information to determine if it's a memory related issue. Have you verified this using the memory pressure graph in Activity Monitor? https://support.apple.com/guide/activity-monitor/check-if-yo...

[skybrian]

It's strange how much people's experiences vary. I'm typing this on an M2 with 8GB and browsing in Chrome works totally fine.

I wonder if it's due to extensions or something like that? Or could it be due to browsing different websites or using different networks?

[art0rz]

Could be, but I got the same results in incognito mode.

[dihydro]

What about with Safari?

[wkat4242]

Yeah it sucks. I use an M1 8GB for work because at work we're not allowed to order CTO models :( We have to order through a reseller that refused to do this. It's very tough to work with (though I'm sure part of the reason is Microsoft's incompetence, their apps perform very poorly and a lot of them are electron crap now).

Apple saying the "our 8GB is equivalent to 16GB" is something they've always done. Even in the PowerPC days they were selling the "megaherz myth" as if their megahertzes were better than intel's.

Of course when they moved to intel Macs received a huge performance boost...

[greatgib]

I have sadly to use one M1 with 8gb and it is really painful.

I think that a normal user will feel the bad effect of the low ram amount but will not really realize where the problem is coming from.

Switching from one app to another is painfully slow. It might take 1, 2 or more seconds to go from one app to another in my opinion.

The os is really greedily evicting app memory once you switch to another window. And reload it when you come back.

I even have the personal hypothesis that Apple is keeping as-is the ergonomically awful Dock because of that. As it is difficult to switch between multiple windows of a single app, you can't easily realize how show you pass from a window to another. Like a magic trick.

[kcartlidge]

> I have sadly to use one M1 with 8gb and it is really painful.

I've been using an 8GB M1 for a couple of years, including for development with C#, Go, Python, and Rails. I've had triple-figures of browser tabs open, Office 2019 Mac, Affinity Designer, and more all running at once. And never in all those years have I experienced sluggishness or slowdown either in-app or during switching.

I don't doubt your experience, this is simply to say that not all observed behaviours can be assumed to be generally applicable.

[etrautmann]

that's crazy to me. I was given a 32GB MBP for work and found it to constantly eat all of the ram and slow down to the point of being useless and require restarts every few days. It took considerable active management to keep more than 4GB free. I upgraded to a higher ram machine now and all of my challenges went away.

My main workflows are python dev and data analysis, some CAD modeling with fusion, and inevitably keeping too many chrome tabs open. Still, with what seemed normal to me, 32 GB didn't cut it, but reading these discussions is crazy making since I seem to be the only one.

[robenkleene]

> It took considerable active management to keep more than 4GB free.

This illustrates ignorance of the basics of how memory management works on macOS (e.g., "free" memory is not a determination of whether you need more RAM) https://support.apple.com/guide/activity-monitor/check-if-yo....

[throwaway89201]

> As it is difficult to switch between multiple windows of a single app

Magic trick: press command + ~ to quickly switch to another window of the same app.

[dzhiurgis]

Wizard trick: map this to four finger trackpad force press (via BetterTouchTool)

[dzhiurgis]

I'm 90% sure your workplace supplied malware ala cloudstrike is causing this.

[wkat4242]

It's more than stingy. It's a calculated bet to make the base model suck so more people will buy the extra memory at Apple's hugely inflated pricing.

[daft_pink]

The real problem is that the default option is the one with mass distribution and if you want it on sale or discounted or on release day, it’s the one you are probably going to end up with that most users have. It’s not only the way you get soaked on the upgrade but the way most users have these low specced devices.

[jacobp100]

Is it swapping to disc, or compressing the memory?

[sudofail]

I thought this as well and bought my wife the 8GB model, thinking it would be enough for her casual use. With nothing open but Chrome or Safari (dozens of tabs, sure) it will lock up and sometimes even hard restart. 8GB is just not enough for anyone anymore.

[acdha]

Either there’s something much bigger running or the hardware is defective. I used an 8GB M1 with VSC, Podman, Slack, etc. for a few years during the pandemic and it was fine. Chrome is a notorious memory hog but even then it was okay so I’d review the installed extensions if it’s that high.

[kcartlidge]

I'd strongly consider returning the machine. This is very far from my normal experience with an 8GB M1 over several years (as a developer as well as a casual user).

[windex]

For light uses maybe even a low cost windows machine would do. No point buying apple.

[gabrielhidasy]

There's never a 'point' buying Apple regardless of light/heavy use. You buy Apple if you think it's the best OS/hardware combo, you don't if you prefer otherwise.

Even light use will benefit from long battery life, nice screens, and quality construction, and a low cost windows machine will not have that. You can get a good windows machine with comparable hardware, but then it's not low cost anymore.

[sigmoid10]

If someone's primary use case is heavy media editing, I wouldn't recommend an Air in the first place. It has its uses for casual people and coders who primarily work in the cloud, but let's not pretend that it is a high end business level laptop.

[pavlov]

I use a 2023 MacBook Air with 24GB memory, and I can’t tell the difference to a MacBook Pro.

Integrated memory on Apple Silicon has made the trade-offs much simpler to understand because there’s no separate GPU and VRAM to think about. The performance difference between Air and Pro models is so small, it’s irrelevant for practically anything.

[sigmoid10]

I have used both. Air and Pro. For working remotely in the cloud on ML, nothing beats the Air. For nearly everything that runs locally and goes above light coding (e.g. modern web dev with endless node modules), the Pro is significantly better.

[aidanlister]

We have bought 200+ macbook pro’s over the years, but when the M3 air came out we started rolling that out to everyone including developers and it has been a huge success. I would only recommend Pro’s for heavy media editing applications right now. We don’t do any AI/ML stuff yet so not sure how that will change the landscape.

[sigmoid10]

I use an Air for ML. It's perfect if you work in the cloud. Locally it's too much hassle. Not just because of the limited power, but because of lack of support for MPS in recent kernels in modern frameworks.