fushuan

I’m necroing the post but I’m happy to tell you that I changed to qobuz a while ago and they definitely have radios now. You can pick a song and create a radio from it. Also, what I think you want I autoplay, where when the playlist finishes it keeps searching similar songs. That works in quobuz now, I listen to the weekly or daily recommendations and it just keeps playing. The recommendations are still not that good since it’s only been a month of not using it daily, but they are getting better.

Deeply would translate from Chinese, maintaining the idioms or minimally adapting them. LLMs take your input and generate a seemingly similar output, which technically says the same thing, but the writing style is completely different.

I write both in English and Spanish all the time, and sometimes I give a pass through a translator or LLM to touch up some emails for work, and the difference in writing is very obvious. One is a translation, the other is what your English buddy wrote after you explained to them what to write. Sometimes I do want that corpo bullshit speech that I can’t come up with natively because the email is for some corporate bastard that will appreciate that vomit, though.

Edit: sorry for the long post,but to expand on the “Deeply is not a translator,but a language model”. Most translators nowadays are language models, NLP was originally developed for translation, although it opened the door for LLMs to exist. What I wanted to say is that not all language models work the same way, and the way language models are used in translation and within LLMs is very different.

Disclaimer: English to Spanish translation is one of the best in the world due to the amount of shared text we have, and the writting style, idioms and such don’t change as much and they would for Chinese, so I understand why they would prefer to format it via an LLM. Still, maybe it was too much.

Nothing alike. Voting is a trust issue, and while we have several trust based systems, it’s very hard to implement a system where even if the users don’t trust the system itself they can trust the result it produces. That’s the issue with electronic voting, trust that is placed in manual vote counting volunteers supervised by representatives of every party is small, because you only need to trust that each box was counted correctly. When you aggregate all those small trusts into a singular system, which even if divided in machines, is still a singular entity, then it’s impossible to trust the centralised system enough, and it must be incredibly robust to prove to every individual that their vote was cast correctly.

I proposed a system like that in my last comment, but I still wouldn’t trust such a system simply because the risk someone phishing the machines on voting day is so high, even if it has close to zero probability, that we can’t risk it.

The whole system is built on trust. With regular voting, we trust that there are enough people counting and validating each voting box that there won’t be much tampering. Each group only has to validate their own box, and so the trust required for it is not that high. We also know that each physical vote introduced in the box is there.

With machine voting, even if you use local machines that give results to humans, which then are inputted in a bigger database for counting, the entity doing the initial count for the district “box”, is singular, and there is no way to manually check unless there’s a bunch of humans watching every anonymous “ballot” of that box are registered correctly. You can’t really do that when the ballot is an “anonymous” input in a machine, there’s no paper trail verifiable by the small trusted human group we used to trust.

Even if there are a bunch of machines, trust is lost. Now imagine the actual useful proposals where you can vote from the phone Yada yada. Trust is completely lost.

The only way to be able to validate is for your vote to have an ID, that you and only you know is yours, and for each vote ID to be public. That way each individual person can validate that their vote is being correctly counted in their district, and trusted individuals can correctly corroborate that the amount of imputed votes makes sense with the amount of people registered to vote in that district.

That puts at risk that people without tech expertise wouldn’t know how to validate or even store their secret ID so others don’t know it’s them, and that we would need to trust the government that they don’t store a database of ID - citizen.

And yeeeeeess, blockchain based ledgers would be good for nonce (secret ID) based voting, but even if that tech works well, socially is still hard to implement as stated before.

Evolve. Or Don’t Look Up.

Criiiiiiingeee

Self righteous self promoting new account that uses AI to write posts for them and is proud about it.

Criiiiiiingeee

Sonar has a functionality that lets you do stuff when conditions happen, I have a script that auto runs when sonarr grabs the releases that searches for my preferred subtitle track and audio track. I can get it after work if you are interested.

I believe Radarr does the same. Also, as others have stated there’s Bazarr for subtitles, but that doesn’t set default subtitle and audio tracks automatically iirc.

If company is gonna lock downs extensions and force SSO on us, making Firefox non-viable (not that I’m allowed to install it anyway)… Fuck it edge we go, they have a Microsoft SSO stack anyway, it’s not like Microsoft isn’t tracking what I’m doing already lmao.

Idk man, why put it? I agree, I down voted the proposal since it was useless. Didn’t go mad posting about it tho.

I understood from the beginning that you are trying to make a “slippery slope” point, but this is open source, each change should be evaluated as is, with what it implies. A local field that isn’t being used in anything doesn’t condition users or Devs to anything that will then make them accommodated and easier to approve an actual invasive feature.

I will agree with the slippery slope argument when they propose a feature that is minimally invasive. This was both useless and 0 invasive.

Edit: actually no, this feature wasn’t useless overall. It was useless for age verification, but great for parental control. The moment a kid doesn’t have root access to the computer, a parent can put whatever age to block the kid from whatever features the parent wants to block them from. Think about it, self enforcing age verification doesn’t give power to governments, it gives it to the root user of the computer, aka parents. It’s something that actually works.

It really doesn’t. You are showing you don’t know how it works.

Webpages can enforce remote verification for sure, that would fuck anyone, Linux included, but a local data file doesn’t leave anything open for the idea I just said.

If you have root access you have complete control of what happens in your local environment. The only way to enforce user verification is to make it remote reliant (just like it’s done in Spain for example, government regulated digital certificates), and then this new field is useless.

It wasn’t a good proposal given that the original intention was compliance in a very useless way, but y’all are going crazy without learning about it.

The first thing on the post you linked is the systemd change which adds a new number field in a completely user controlled local environment where they can write anything they want.

Oh nooooo… ಠ_ಠ

Are you referring to the userd field? That would mean nothing if the user has total control on what can be written there.

Whatever device based verification those websites or electron apps were communicating with can be spoofed in a system where you have complete control.

Games are cracked in weeks at most, don’t you think that whatever secure communication is established won’t be cracked lightning fast by the whole FOSS community? Once the “secure communication” between local apps is broken, a third package can mitm that shit easily. It’s a local environment.