Yep. And you can curate them. For example https://piefed.social/f/fediversevideos

Its a mix of communities, peertube channels, and mastodon accounts attached to videos.

Yep. Thats what ive seen on both my instances. Piefed has had a lot of new features. Lemmy has had stable dev work.

Just take out the battery of any phone like a fairphone and most will work.

WOOOOOOO!

Dang…

Hope they come back soon. ;(

Heh coo.

Any idea if SDF is coming back? Loved that lemmy instance and the community around it.

Essentailly. Theres a couple of peertube instances that had that issue if i recall. Filters started to go up a coupke of years ago. Admins can see it pretty easy in the logs most of the time once its pointed out. Its a timing thing. Some software is beyter at it than others.

Same fantastic one liners.

Good questions. Its been a bit since ive looked at lemmys code so im going to dive in again.

Here is the docs for the votes themselves. https://github.com/LemmyNet/lemmy-docs/blob/main/src/users/03-votes-and-ranking.md

You can see this in action here: https://lemvotes.org/ if I recall.

I know the vote model is here . You can see the logic here. It checks to see if the post is actually available in the community with other functions and other such checks. It doesnt ask any lemmy instance for totals (unless theres something im not aware of), it tallies based on real time data. Which can also include fediverse actions(not necessarily lemmy/piefed/mbin/etc… but also mastodon, gotosocial, and other such services).

One of the things piefed does a bit differently is it bunches up votes for sending to other instances. And there is a backfill operation when puling from new communities or users. I would argue its a bit better than lemmys system at least from a technical perspective. But its a VERY minor one. Its nice we have multiple ways of getting to the same result-ish.

Hope that helps! I dont know everything but I know a little bit.

It would be VERY difficult. I have no idea what im looking for other than the delay or block. And im not seeing anything today. If you can give me an example, I can check in and reproduce. Otherwise its nettle/haystack.

It might very well be a bug.

I can see it being frustrating. But thats what community software is. The good news is that even just reporting the issue can help us all out!

Yep i can. Ive helped out on both lemmy, piefed, mastodon, gotosocial, peertube and others. Little things mind but yeah its my day job.

Theres a queue that is part of different instances. What can happen is posts/comments get backed up and the queue takes them all in. When an outage or slowdown happens it can bog down certain servers. Threadiverse (piefes/lemmy/mbin/etc…) up/down votes, comments, posts, etc… it gets slower as more instances connect to others. Its unfortunate but its part of the fediverse protocol. Piefed has some advances there, but it can still get bogged down sometimes.

Im not saying thats definitively what it is, but ive been on the other side of this phenomenon a couple of times when my lemmy and then later piefed instance had a new instance sync with it. Lemmy.world has had this issue for a while since its one of the biggest.

The way to test is to sent the instance in quesion a POST request with a comment OR use a known good instance (like piefed) and check the logs. It usually comes back with either a 200 immediately (which is good) or theres a HUGE delay.

Most of these instances are very much volunteer efforts.

Can you elaborate? Arent blocks public via the admin logs in lemmy? In addition that info can be pulled from the api from any instance/fediverse protocol.

I can reproduce of need be since i have my own lemmy and piefed instance and have over 20 years of software development experience.

I dont understand. Can you point to where in the code it has this? The algorithm/stack is open to anyone.

How would that work?

We are all on different instances. With different algorithms, technology, and differing filters. The code doesnt support what your saying.

My research was AI. Using a certain ai algorithm to find cancer tumors. The new ai funding is helping out adjacent projects.

Llms im still kinda iffy about since the tech requires a lot of copyrighted works just to get mediocure results.

AI is a HUGE field. But what people talk about is the llm side of things.

Yeah im seeing it lol.

The community is pretty split. I know a lot of people are going to think the accoints are bots. Maybe they are.

But ive met people in real life that truely believe in llm ai solving all thier problems. Its not true bur thats what they believe.

Datacenters are going to heat up the us so much. Thats exceptionally bad for the people and environment around them.

Dank pods / James is great. My wife and I like watching during lunch.

nice. Ive thought about getting a smart watch but specifically for blood pressure. But ive seen a lot of reviews saying its not accredited. Which kinda sucks.