I'm running a lot of tools in parallel and you need memory especially when you're doing memory tracing and debugging, which is kind of obvious. I mean, when you load 4 GB memory dumps, you tend to run out of memory pretty fast ;-) Couple this with a compiler that watches for file changes and constantly recompiles stuff in the background, runs tests and so on and with a full fledged IDE, like IntelliJ IDEA, with a browser loaded with stuff and things get painful.

I'm working on a MacBook Pro with 8 GB of RAM and it constantly swaps to disk. Thank heavens I have an SSD, because otherwise it would have been unusable.

As to your response, dude, this attitude of "you're holding it wrong" is iconic of Apple. I mean I don't know what people are talking about, as Steve's spirit is very much alive at Apple and within its user base and I never liked it.

> I'm working on a MacBook Pro with 8 GB of RAM and it constantly swaps to disk. Thank heavens I have an SSD, because otherwise it would have been unusable.

Don't get me wrong, but if you're doing this kind of serious development, perhaps a laptop isn't the best choice? The iMac 5K takes 64GB these days.

Perhaps I'm just too old and can't see why a laptop is a superior development environment compared to a desktop with decent 27" display, mouse and keyboard... Are you traveling much?

Because then I'd have to buy an iMac 5K for home and one for the office. And sync all my files and preferences between the two.

Or, you could buy a beefy desktop at work and use Apple Remote Desktop / VNC on your laptop at home.

If you've ever used Remote Desktop / VNC, you know that it just doesn't work. The latency of the network communication is too great within your own local network over Wifi, enough as to be a pain in the ass for work, let alone when communicating over the Internet. And you'd better not be on a metered Internet connection, because these high-resolution screens will produce data like crazy.

I know it sounds cool in theory, but it just doesn't work, at all. Basically Remote Desktop / VNC is only useful for short sessions and/or debugging, not for work, unless you're in text mode with SSH, vim, screen, tools like that ;-)

>If you've ever used Remote Desktop / VNC, you know that it just doesn't work

I wrote a DOS VNC client in 1999 for my ThinkPad 720 (to access my Sun workstation at work from home), so my mileage varies. ;-)

We had lower screen resolutions and color depths then, but also slower connections. I've used VNC quite a bit over the years but admittedly, for working from home I used ssh (because it was sufficient and I use vi). FWIW, practicality greatly depends on your type of work. Typing text in an IDE and running the occasional debug session should be no problem due to the way screen updates are transmitted, anything graphics-heavy will be frustrating.

Personally I have two development 'modes', depending on what kind of project I'm working on. One which I can easily do on something like an MBA with 4 GB of RAM and one which I really want at least 32 GB of RAM and decent Nvidia card. The new MBP falls in between these two modes and as such I can't see the niche it would fill for me.

Wow, reading your comment was just like reading my own words. I am exactly same (minus Nvidia part because I don't do intensive gpu stuff).

MacBook air is my favorite device. Giving presentation, developing small apps while giving presentation, pair it with 32gb i7 fedora desktop for serious projects. This combination is my favorite.

Now it seems they are discontinuing MacBook air , I will be happy with my 8gb air version for next 3-4 year. But after that I don't know what can come close to replacing it. Maybe surface nook.

The new MBP 13" is smaller than the Air, so the only issue is the price?

Although I didn't know that and it is wonderful, but yes price is a little bit annoying. MacBook air was perfect for me.

Sorry for a little bit typo. I meant surface the tablet ones.not book.(book is so pricey)

Containers only get you so far. Windows e.g. tends to be a real memory hog, especially older versions (Windows Update alone can need 4+ GB RAM… per VM).

And if you can't shove everything into The Cloud® because your country is lagging behind on decent internet connections and 1 MBit uplink is the best you can get… well.

People that need more memory are most likely to be the ones commenting regarding memory. I suspect that the majority of developers aren't going to need more memory, at least not for 99% of their job.

Also no doubt people want a bump on the specs from the last generation, even if it isn't needed.

Also if you buy now, you want your MBP to last a while, 16GB is the new 8GB - it's fine now, but it may hit a limit in a handful of years.

Yeah, I kinda agree. But it really depends on use case though. I do data analysis and I have a beefy server that I log on to for heavy duty tasks, so I'll probably not use my macbook for anything that requires more than 16gb.

I have 32 GB of RAM on my main computer and that actually saves me from having multiple computers as I use a couple of VM's (at least one running all the time) and the typical memory usage is between 18 and 28 GB. I could probably live with 24 GB most of the time, but not with 16 GB or less. I have 8 GB on my laptop in the office and I use it only for email and browsing and to remote desktop to bigger boxes. But my laptop does not have "Pro" in the name and it's dirt cheap.

Chrome alone takes 12G currently. I have multiple opened Chrome windows with about a dozen tabs each. One window for youtube and other sites that tend to make sounds, one for each client project I'm currently working on, and one for each personal project.

Filling up 16G is pretty easy if you seldomly shut down the PC.

I never managed to open that much tabs, maybe I am doing something wrong.

Having a single GUI-supported VM with 4 gigs is a no-go. You must also use a very light IDE, vim/Sublime/etc.

I'd like to run cloudera on it :)

> Do you guys run kubernetes on it, or why you need more memory?

Yes we do. Though 16GB is well enough for that.

It depends what do you run in your pods.

There's a significant speed increase going from 4GB to 8GB RAM.