I have a very cynical reason. If you look at what most religions say about it (against), you have to wonder why they all agree on it and it seems to me that if you off yourself, you’re not supporting the team. When there weren’t many humans, you really needed a bunch of team players on your religion making more babies, and the dead ones can’t carry out your crusades.

Now we put capital above religion, but it’s the same thing: we need workers for our factories. We need babies to become workers for our factories. Dead people can’t make cars or babies.

I don’t care for a lot of what Apple does, but there’s no denying they understand how to make money–and how to avoid losing it.

I don’t see how it statistically benefits the company. Whether I sell you the right thing, or the wrong thing, I still sold you something. So why not try to make it the right thing so I come back?

Thirded. Immich has no right to be as good as it is after such a short time. Completely took down my google photos, finally, and I still have face recognition, word search and automatic backup from my phone.

My favorite thing I’ve done with hass is put a color-changing light bulb by my front door. It’s connected to the weather forecast. I know what the weather will be at a glance without a website or going outside. (Where I live, it’s not always obvious when I’m gonna get rained on.)

• uptime kuma

Not exactly the same. First of all, a new company hasn’t formed yet and if it does it won’t retain as many of the original staff because it doesn’t have the same momentum as one formed through a legal separation of the companies while everyone is still employed.

It also can’t retain the rights to Stray which would have provided some funding. And it doesn’t retain the rights to whatever projects they had in development, so they won’t have anything to work on for a while.

That legal spinning off was actually pretty important.

Mostly it’s just CYA for google since cycling is more dangerous than driving (due to the people driving), so there’s more surface area for them to get sued.

But yeah

• turns and crossings that look safe on a map don’t have very much data on whether they’re actually safe, because google has a thousand times as much information about drivers than cyclists.
• google sometimes suggests routes that can’t be traversed, legally or at all, by a bike. Same reason.
• sometimes google suggests avoiding something a bike doesn’t actually have to worry about. This is actually the category of error I see the most: google sends you around something when you could simply walk your bike through it, or ride through it, because you’re not a car.

Linus is the leader of the kernel project. As a leader, it’s his job to get the maintainers to agree. It’s not Rust’s job to make the C devs stop bullying them.

If Linus thinks Rust is a good direction, he should show it by actually standing up to Ted and developers like him and making them behave.

If he doesn’t think it’s a good direction, he should say that too, so the remaining Rust devs can stop wasting time on the project.

When someone in a niche part of the project steps down like this, that’s a problem with the top-level leadership. Linus’ record on leadership is… mixed. Trending in a good direction the last few years, but this makes me wonder. He can still save this, but he has to want to.

Bcachefs has all of this. And it’s supposed to be faster than ZFS and btrfs. In a few years it can really be the golden Linux filesystem recommended for everybody

ngl, the number of mainline Linux filesystems I’ve heard this about. ext2, ext3, btrfs, reiserfs, …

tbh I don’t even know why I should care. I understand all the features you mentioned and why they would be good, but i don’t have them today, and I’m fine. Any problem extant in the current filesystems is a problem I’ve already solved, or I wouldn’t be using Linux. Maybe someday, the filesystem will make new installations 10% better, but rn I don’t care.

Specifically Italians, at that.

Heck yeah hook us up

Podman is not yet ready for mainstream, in my experience

My experience varies wildly from yours, so please don’t take this bit as gospel.

Have yet to find a container that doesn’t work perfectly well in podman. The options may not be the same. Most issues I’ve found with running containers boil down to things that would be equally a problem in docker. A sample:

• “rootless” containers are hard to configure. It can almost always be fixed with “–privileged” or some combination of permission flags. This would be equally true for docker; the only meaningful difference is podman tries to push everything into rootless. You don’t have to.
• network filesystems cause headaches, especially smbfs + sqlite app. I’ve had to use NFS or ext4 inside a network-mounted image for some apps. This problem is identical for docker.
• container networking–for specific cases–needs to managed carefully. These cases are identical for docker.

And that’s it. I generally run things once from the podman command line, then use podlet to create a quadlet out of that configuration, something you can’t do with docker. If you are having any trouble with running containers under podman, try the --privileged shortcut, see that it works, and then double back if you think you really need rootless.

I haven’t deployed Cloudflare but I’ve deployed Tailscale, which has many similarities to the CF tunnel.

• Is the tunnel solution appropriate for Jellyfin?

I assume you’re talking about speed/performance here. The overhead added by establishing the connection is mostly just once at the connection phase, and it’s not much. In the case of Tailscale there’s additional wireguard encryption overhead for active connections, but it remains fast enough for high-bandwidth video streams. (I download torrents over wireguard, and they download much faster than realtime.) Cloudflare’s solution is only adding encryption in the form of TLS to their edge. Everything these days uses TLS, you don’t have to sweat that performance-wise.

(You might want to sweat a little over the fact that cloudflare terminates TLS itself, meaning your data is transiting its network without encryption. Depending on your use case that might be okay.)

• I suppose it’s OK for vaultwarden as there isnt much data being transfered?

Performance wise, vaultwarden won’t care at all. But please note the above caveat about cloudflare and be sure you really want your vaultwarden TLS terminated by Cloudflare.

• Would it be better to run nginx proxy manager for everything or can I run both of the solutions?

There’s no conflict between the two technologies. A reverse proxy like nginx or caddy can run quite happily inside your network, fronting all of your homelab applications; this is how I do it, with caddy. Think of a reverse proxy as just a special website that branches out to every other website. With that model in mind, the tunnel is providing access to the reverse proxy, which is providing access to everything else on its own. This is what I’m doing with tailscale and caddy.

• General recs

Consider tailscale? Especially if you’re using vaultwarden from outside your home network. There are ways to set it up like cloudflare, but the usual way is to install tailscale on the devices you are going to use to access your network. Either way it’s fully encrypted in transit through tailscale’s network.

I hope you’ve turned on 2FA.

The objection about a “finite planet” is about capitalism, not currency. A 100% communist system can still have fiat currency and function perfectly well, the two aren’t even related.

It’s capitalism you don’t like, not money.

Wouldn’t even take a month, just prepay for those reserved instances.

Thanks! I’ll try this and report back. This sounds like a version of (#1) - merge accounts.

Home assistant’s main use case is showing you where your house is on a single map, though. Not sure how immich works, but if it’s one tile per photo with location data, that would be a MUCH bigger ask.