I really didn't enjoy the cheap plastic construction of the Moonlander. I had two of them for home and work. I even modded a mousepad onto the wristrests to make it more comftable.
But in the end the housing being out of plastic, it creaked, wobbled and just was not satisfying to type on.
I came from premium mechnical keyboards with solid steel or aluminum construction.
I ended up with the Neo Ergo, a middle ground. Not as ergonomic, but solid feel, no plastic and great looks too.
You can keep "columnar" data in a row based database like postgres, it's just more expensive. But with little data that's fine and reduces infrastructural complexity. When you reach too much data it gets to a point where you then actually want to use the correct database for your usecase.
I'll remember it and email you when I try it again.
Last time I tried all of orbstack froze and I had to restart my whole mac to fix it. But you also did some recent releases that fix issues related to freezing up, so maybe it was unrelated.
Thanks for the great software! Happy enterprise customer
I just wish bind mounts would be more performant/native. I get that this is probably impossible, and probably also sucks on Linux, haven't tried.
But like having containers that need file watchers like vite dev server, or frankenphp in watch mode will overload OrbStack real quick since It seems to fallback to polling instead of listening to fs events.
So I'm stuck running vite dev servers and the like on the host.
I agree. This thread is so negative, seeing everything black and white. 99% of the users of these tools are never going to bother trying to figure out how to remove synthId. And then for the small percentage that do bother, they would have the knowledge to usage image models that don't use watermarking in the first place.
No single person can draw that line, that's what Courts and Laws are for. And some of the industries play more dirty and try to manipulate that due process, others failed.
But that's what we have, it's never black & white. Always a process and always evolving.
I think this is mostly to blame on Grafana. Every dashboard ships with $__interval by default and every query uses that by default. I never understood the value of that. It makes the graphs look pretty I guess, but looking at the values becomes a little less useful because nothing is compareable to another not even to yesterday when you last looked at the dashboard.
Personally I always setup a variable $interval that can be set to 1m, 5m, 15m etc and use that in all my queries.
The integration of the ed25519-sk keys is just so easy and similar to normal ssh keys, so the upgrade is way easier.
You just need to tighten your sshd config, you can even add a "touch required" of the Yubikey to the sshd config. Has been in debian stable since like 11 I think?
So it's super friendly to integrate and very secure, as you need to physically be on your pc, have your yubikey and have your exact pc. So that's a lot of factors.
The amount of traffic that must go through the docker hub, must be insane, the images go into the gigabytes and everyone expects it to be super fast.
Cloudflare is one of the few companies who can handle that for relatively cheap, Docker could not just "run their own" and have it even be compareable.