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

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”.

creating a label and checking the skip invoice box

That works great too, specially if you want to use less foolproof filters. Or even a mix of both strategies.

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!

If you want, you could use GMail filters to delete those emails automatically. Here’s how:

1. click the engine button (settings), then “see all settings”, then “filters and blocked addresses”.
2. click “create a new filter”. Add “top of Google search” to the field “has the words”, leave other fields blank.
3. click “create filter”, then check the “delete it” box, then “create filter” again.
4. repeat steps 2-3 for other shit that SEO spam is likely to mention.

Important: never use as a filter anything that legitimate users might reasonably say. Only things that you’re fairly certain to come from a spammer.

EDIT: I repeated two steps without noticing it. My bad.

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/

It’s less complicated than it looks like. The text is just a poorly written mess, full of options (Fedora vs. Ubuntu, repo vs. no repo, stable vs. beta), and they’re explaining how to do this through the terminal alone because the interface that you have might be different from what they expect. And because copy-pasting commands is faster.

Can’t I just download a file and install it? I’m on Ubuntu.

Yes, you can! In fact, the instructions include this option; it’s under “Installing the app without the Mullvad repository”. It’s a bad idea though; then you don’t get automatic updates.

A better way to do this is to tell your system “I want software from this repository”, so each time that they make a new version of the program, yours get updated.

but I have no idea what I’m doing here.

I’ll copy-paste their commands to do so, and explain what each does.

sudo curl -fsSLo /usr/share/keyrings/mullvad-keyring.asc https://repository.mullvad.net/deb/mullvad-keyring.asc
echo "deb [signed-by=/usr/share/keyrings/mullvad-keyring.asc arch=$( dpkg --print-architecture )] https://repository.mullvad.net/deb/stable $(lsb_release -cs) main" | sudo tee /etc/apt/sources.list.d/mullvad.list
sudo apt update
sudo apt install mullvad-vpn

The first command boils down to “download this keyring from the internet”. The keyring is a necessary file to know if you’re actually getting your software from Mullvad instead of PoopySoxHaxxor69. If you wanted, you could do it manually, and then move to the /usr/share/keyrings directory, but… it’s more work, come on.

The second command tells your system that you want software from repository.mullvad.net. I don’t use Ubuntu but there’s probably some GUI to do it for you.

The third command boils down to “hey, Ubuntu, update the list of packages for me”.

The fourth one installs the software.

The first response seems reasonable for me; it’s informative and replying to an ambiguous comment, as you can’t quite know if “isn’t there” refers to his individual needs or in general.

The second response is however passive aggressive garbage. Fl4ppers clarified that he was talking about his individual needs; notjustforhackers failed to take it into account, and his response sounds a lot like “I’m just sayin lol lmao… you liar”.

Ghostery does work as an ad blocker too, and advertises (eh) itself as such.

Plus, for corporate nowadays, tracking and advertisement are two steps of the same process.

As the other poster said, it’s on F-Droid. It’s mostly a Youtube frontend, although it accesses SoundCloud and PeerTube too. And more importantly, it doesn’t download advertisements from YT, so you don’t even need an ad blocker when using it.

The ad blockers in question were AdGuard and Ghostery, acc. to the article.
uBlock Origin is still working fine for me. And NewPipe in the cell phone.

Frankly? I’d rather stick to a normal package manager too, if available. But the alias trick is useful in a pinch, if you must use a flatpak.

As a local fix, you could set up an alias. Open .bashrc and add the following line: alias firefox="flatpak run org.mozilla.firefox"