> why does it even matter if it's faster?

Have you considered that plenty of people don’t need VirtualBox to write code?

Maybe you should be questioning why you need VirtualBox to write code

Security, provisioning, dependency management, infosec, etc. the list of reasons for someone to develop in a VM is endless.

yes. also vbox is not the only way to run VMs

Parallels supports linux VMs. https://kb.parallels.com/128445

You could use that now or wait for virtual box to catch up (I assume).

I'm writing this comment from Ubuntu running on my M1 Macbook Pro, so...?

How is the GPU performance?

No idea really, not sure I've done anything GPU-heavy on Ubuntu.

You can run Linux with UTM.

Can I keep my existing vagrant / Ansible setup? Or, basically, what would I need to use in order to provision the instances?

You can keep your ansible. Does vagrant support alternate providers for running VM's? Then yes, you can keep your vagrant too.

It looks like Vagrant doesn't work yet with UTM: https://github.com/hashicorp/vagrant/issues/12518

Conceptually I'd love one of these machine, but realistically if I can't develop locally in an Ubuntu VM then I might as well just wait a year or two to buy one until that's possible. I'm not going to use Docker, and realistically I have zero interest in spending multiple weeks rewriting all of my provisioning scripts.

Vagrant has a QEMU driver, which is what UTM uses anyway. I can't say it will do what you need the way you want but I suspect it's possible to get there.

QEMU can virtualise arm and emulate x86.

Thanks, I'll check this out! I just need to run Python on Ubuntu, so hopefully that isn't too crazy of a use case.

You can always use docker

Docker doesn’t really work on Mac. Like it technically runs, but just barely. Because it uses containerization inside virtualization, it pegs the CPU just to run a hello world app and completely drains the battery in 30 min. Oh and the fans sound like an airplane taking off.

I don’t know anyone at work who would agree with this. Docker is fine on macOS, honestly always has been. Sure, it will show up in Activity Monitor, but so will any other software that is actually doing something. I have never seen my CPU peg because of Docker, and I regularly run large stacks, use Earthly for all my builds, etc. The overhead of the VM + containerd is truly not that large.

CPU % in top and similar tools isn't a great proxy for energy usage anyway, it's a lot more complicated than that.

Docker is terrible on Macs, but I used to run Docker+vscode+Firefox+Brave+Slack+misc. I'd run into issues on my 2015 MBP from time to time, but on the 2019 models it was fine.

gulp.watch would often eat an entire core, and stuff in Docker ran slower than on Linux, but my laptop could keep up. vscode didn't lag and Firefox ran just fine—all with 5+ Docker containers running.

I have no idea what you are talking about. I have about 7 containers running the full work app and I can't hear any fan noise and the laptop is running cold.

This thing is running Rails, mysql, elastic search, redis, mongodb, and workers inside docker. As well as nodejs, vscode, a handful of electron apps all at the same time and the laptop is cold.

Not my experience at all, you must have something weird in your setup. For a full dev environment I'm running containers with postgres, cockroachdb, several ruby apps, kafka, some JVM based stuff, and clickhouse and it does hit battery life a bit but certainly not pegging the CPU, actually it's using less than Slack...

Not my experience either on a fanless M1 Air. I can run lots of containers without it perceptibly heating up or slowing down. Battery life does suffer but I can still usually get at least 4-5h of solid programming like this.