Removed by mod

Sorry for the double reply. Here’s a practical idea: what if the mods of this comm contacted lemmy.ml’s admins? Ideally doing two things:

1. Clarifying that the instance in question does not have child sexual abuse material, and requesting users to be allowed to link it.
2. Expressing desire to migrate !anime@lemmy.ml to the instance in question, and highlighting that this is convenient for both sides of the matter.

Among the admins I think that Nutomic would be the best to contact, given the github thread.

You’re talking about your thread about Mahou Shoujo ni Akogarete, right? It’s still in the modlog for me, even in private mode. I don’t think that they removed the entry.

Another important detail is that Digg v4 pissed off most of the userbase, so the impact was pretty much immediate. Reddit APIcalypse pissed off only power users instead; the impact will only come off later (sadly likely past IPO).

Lunix sucks so much that it got stuck into the version 2 for years.

Removed by mod

In my opinion, the migration is sensible because:

1. You need to know a topic decently enough to be able to moderate a community about it. And yet it’s unreasonable to expect the admins of lemmy.ml - a community about free/open source and privacy - to know about anime.
2. There’s no inherent reason why this comm is the largest anime comm in Lemmy. It’s simply that this comm is three years old, from a time when “Lemmy” was mostly just lemmy.ml and lemmygrad.ml, there was no other place in Lemmy to discuss anime.
3. Lemmy as a whole benefits from redirecting traffic from larger instances to smaller ones. Specially instances unrelated to politics and tech. In fact, lemmy.ml’s admins asked users to use other instances not too long ago.

For me it says “admins”. I don’t know if it’s because I’m a mod of a few other lemmy.ml comms, or if it’s because of the different backend versions (SJW uses 0.18.5, lemmy.ml uses 0.19.1).

For context, I encourage people to check this discussion in the “join Lemmy” site github. Have in mind that both of the Lemmy developers in that discussion are also admins of the lemmy.ml instance, and they clearly disagree if the instance in question should be considered as “hosting CSAM” or not.

I think that this community should migrate, but for a different reason: topic.

The topic of lemmy.ml is privacy and free-as-freedom software. Most other content here is off-topic, including anime. It was fine when it was just lemmy.ml and lemmygrad.ml, as you’d have no other place to discuss anime in Lemmy, but now the situation is different.

And ideally, communities should be managed/moderated/administered by people who know well the topic of the community.

Aaaaah. I really, really wanted to complain about the excessive amount of keys.

(My comment above is partially a joke - don’t take it too seriously. Even if a new key was added it would be a bit more clutter, but not that big of a deal.)

The source that I’ve linked mentions semantic embedding; so does further literature on the internet. However, the operations are still being performed with the vectors resulting from the tokens themselves, with said embedding playing a secondary role.

This is evident for example through excerpts like

The token embeddings map a token ID to a fixed-size vector with some semantic meaning of the tokens. These brings some interesting properties: similar tokens will have a similar embedding (in other words, calculating the cosine similarity between two embeddings will give us a good idea of how similar the tokens are).

Emphasis mine. A similar conclusion (that the LLM is still handling the tokens, not their meaning) can be reached by analysing the hallucinations that your typical LLM bot outputs, and asking why that hallu is there.

What I’m proposing is deeper than that. It’s to use the input tokens (i.e. morphemes) only to retrieve the sememes (units of meaning; further info here) that they’re conveying, then discard the tokens themselves, and perform the operations solely on the sememes. Then for the output you translate the sememes obtained by the transformer into morphemes=tokens again.

I believe that this would have two big benefits:

1. The amount of data necessary to “train” the LLM will decrease. Perhaps by orders of magnitude.
2. A major type of hallucination will go away: self-contradiction (for example: states that A exists, then that A doesn’t exist).

And it might be an additional layer, but the whole approach is considerably simpler than what’s being done currently - pretending that the tokens themselves have some intrinsic value, then playing whack-a-mole with situations where the token and the contextually assigned value (by the human using the LLM) differ.

[This could even go deeper, handling a pragmatic layer beyond the tokens/morphemes and the units of meaning/sememes. It would be closer to what @njordomir@lemmy.world understood from my other comment, as it would then deal with the intent of the utterance.]

Not quite. I’m focusing on chatbots like Bard, ChatGPT and the likes, and their technology (LLM, or large language model).

At the core those LLMs work like this: they pick words, split them into “tokens”, and then perform a few operations on those tokens, across multiple layers. But at the end of the day they still work with the words themselves, not with the meaning being encoded by those words.

What I want is an LLM that assigns multiple meanings for those words, and performs the operations above on the meaning itself. In other words the LLM would actually understand you, not just chain words.

Complexity does not mean sophistication when it comes to AI and never has and to treat it as such is just a forceful way to make your ideas come true without putting in the real effort.

It’s a bit off-topic, but what I really want is a language model that assigns semantic values to the tokens, and handles those values instead of directly working with the tokens themselves. That would be probably far less complex than current state-of-art LLMs, but way more sophisticated, and require far less data for “training”.

Oh “great”, more crap between Ctrl and Alt.

[Grumpy grandpa] In my times, the space row only had five keys! And we did more than those youngsters do with eight, now nine keys!

Ah, got it. My bad. Yeah, not providing anything is even lazier, and unlike “lazy” bash scripts it leaves the user clueless.

I like them, even for software installation. Partially because they’re lazy - it takes almost no effort to write a bash script that will solve a problem like this.

That said a flatpak (like you proposed) would look far more polished, indeed.

Frankly in this case even a simple bash script would do the trick. Have it check your distro, version, and architecture; if you got curl and stuff like this; then ask you if you want the stable or beta version of the software. Then based on this info it adds Mullvad to your repositories and automatically install it.

You’re welcome.

I think that people being jerks take for granted how confusing this might be, if you’re new; we (people in general) tend to take vocab that we already know for granted, as well as solutions for small problems. …except that it doesn’t work when you’re starting out, and we all need to start out somewhere, right.

You have two options: install curl (check @TrickDacy@lemmy.world’s comment) or do it manually. Installing curl is the easiest.

If you want to do it the hard way (without the terminal), here’s how:

1. Download the file https://repository.mullvad.net/deb/mullvad-keyring.asc from your web browser.
2. Open your file browser as administrator. There’s probably some link for that in the Menu.
3. Move the file that you just downloaded to the directory /usr/share/keyrings/