Ask HN: How do you like developing on an M1 Mac so far?

Up until it died, the MacBook was super nice to work on, super fast in waking from sleep, and pleasant to use in all lighting conditions except direct sunlight obliquely incident to the screen (in which case the sunlight made all the dust and fingerprints on the screen extremely bright and opaque).

Obviously I can't run my intel-based VMs, but I was able to run all my web development toolchains: Django, Vuex/Flask, PHP, Postgres/MySQL/SQLite.

I also play games: mostly KSP. The 8GB RAM isn't enough for a small KSP save (a dozen flights, a few hundred parts), and I had to restart KSP every couple of hours due to memory pressure.

Best computer I've ever owned, but I've already killed one inside two weeks of owning it and the replacement is going out the same way: something in my environment or in the way I use it caused the M1 to kill itself.

It starts off as the occasional panic/reboot. Continues to manifest as any activity that requires touching the disk ends up locking the computer up for tens of seconds (spinning beachball cursor, but you can't move the cursor with trackpad or external mouse). Ultimately the MacBook just dies in the middle of whatever it was you were doing, and nothing you or the Apple Genius can do will convince it to boot again.

sshuttle is definitely no efficiency daemon - however, it only requires being able to ssh to the other side, amd run Python there, and that’s it - whereas for Wireguard you need root on the other side, a sufficiently new kernel (or building an out of kernel module) - it’s way better if you can use it but IME it isn’t “almost as easy” - sshuttle usually just works if you can ssh.