Yoruio

I just realized that this is the homeassistant community and not something more generic.

Specifically for homeassistant, Wireguard should be fine, unless you plan to do some more advanced stuff like use Alexa without a nabu casa subscription.

The times where you need access to your HA instance without being able to connect to Wireguard should be pretty limited.

To use Wireguard, you need to:

1. provision a client tunnel for every device, or at least every person who needs to access your network
2. have Wireguard downloaded and installed on every device, with the tunnels all imported.

Basically, Wireguard works really well for services that only you use, on your own devices. You set it up once per device, and you have access to every service you host on your network.

For the DuckDNS / reverse proxy route, you need to configure the reverse proxy for every service you want to expose, but don’t need to configure anything on the end user’s device.

For Jellyfin, since I have users that are not me, it is impractical to expect them to go through all the hoops to get Wireguard running just to watch some movie or tv show. I also don’t want to make new Wireguard client tunnels for every single friend that I add to my jellyfin server. This also means I can access jellyfin on devices that aren’t my own such as a friend’s TV.

For immich, my phone is a bit wonky with keeping Wireguard connected in the background, and I just don’t want to worry about if I’m connected to my vpn just so my photos will get backed up.

Different services for different use cases.

I use nginx reverse proxy behind Duckdns for anything that requires public access, or that I use very frequently, like jellyfin or immich

I use Wireguard for everything else, to expose as little as possible.

If anything, I would say that Duckdns is harder to setup than Wireguard. You will need something like nginx reverse proxy if you want to host multiple services, and also deal with SSL certificates.

you posted this on a open source platform that someone bothered to create.

In the US, publically traded companies have a legal obligation to make as much money for their shareholders as legally possible (See Ford getting sued by shareholders after giving workers raises). It would be borderline illegal for a company to adjust their algorithm in a way that makes them less competitive.

This needs to be regulated by government, not the companies themselves. Thay would mean that the companies would be forced to all change their algorithms at the same time, and not impact their competitiveness.

So the government going after tiktok is a good first step, IF it does the same thing to Facebook / instagram / YouTube / snapchat. But I’m betting it won’t be because those companies spend an absurd amount of money on lobbying.

this is just how capitalism works - you have to appeal to your audience more than your competition, and guess which kind of content teenagers want to watch more. Hell, even adults want fun content as opposed to educational content.

they’re not willingly selling a worse society for profit, that’s just the only way to stay competitive.

any platform that pushes educational content in North America would just not get any customers and go bankrupt.

edit: there’s plenty of educational video platforms out there, like Khan academy. Try and get your kids to scroll through that during their free time instead, I bet they won’t.

my chrome even sometimes gives me a winky face “;)”

Carissa

Checking in with > 200 tabs in sideberry

I don’t think a OS should ever be LESS open about what a user can do. It should be on the user to do their due diligence and have high availability systems setup.

Only reason Linux wasn’t affected as much was luck. this could just as easily have happened to Linux systems if the broken update targetted Linux.

We (this community especially) criticize windows for not being more open like Linux, and all of a sudden we’re saying it should’ve been more like Apple?

If you had a Samsung fridge, and you willingly put a bomb in the fridge, would you blame Samsung when your fridge explodes?

Microsoft gives you the freedom to install software that runs with the same level of privilege as the kernel itself. You’re the one that chose to install defective software, and then give it kernel level permissions. You put a bomb in your computer and now you’re blaming Microsoft after the bomb exploded.

Microsoft didn’t make the decision to allow the faulty input, the person who installed the software did, when they gave it permission to run in kernel mode.

I don’t remember much about plex photos, but facial (and object) recognition, photo map, easy sharing through albums (without the other person needing an account), and being open source are some features I imagine plex photos does not have.

it seems barebones still, because it is a very young app, and the UI is not great, especially on mobile.

It is the best replacement for Google photos that I have seen though.

+1 to immich

However, because of the fast dev cycle, it has a lot of breaking changes, and needs regular maintenance (most notably for me, postgres docker changes. Especially if you are not giving it it’s own postgres instance and using their provided docker compose.)

You could stay pegged to a single version, but the mobile app also doesn’t have full backwards compatibility with server versions, which results in a slew of other problems (how do I do a fresh install of an older app version on a new device?)

But if you are willing to keep up and perform semi-regular maintenance, immich is great, and the rapid dev cycle means more new features faster!

straight to jail

netlify is amazing for static sites.

The inflation is inflating

Using mugshots to train AI without consent feels illegal. Plus, it wouldn’t even make a very good training set, as the AI would only be able to identify perfectly straight images shot in ideal lighting conditions.

I live in Canada.

Was this AI trained on an unbalanced dataset (only black folks?)

It’s probably the opposite. the AI was likely trained on a dataset of mostly white people, and thus more easily able to distinguish between white people.

It’s a problem in ML that has been seen before, especially for companies based in the US where it is just easier to find a large amount of white people as opposed to people of other skin colors.

It’s really not dissimilar to how people work either, humans are generally more able to distinguish between two people who are races that they grew up with. You’ll make more mistakes when trying to identify people of races you aren’t as familiar with too.

The problem is when the police use these tools as an authoritative matching algorithm.

like ICBMs and freeze dried ice cream sandwiches!