You have 128GB of RAM

Who do you think I am, a multimilionaire?

…and despite that people are upvoting, because who cares about facts, let’s just hate a thing that I don’t like

The main advantage of fediverse over, let’s say X, is that you can change server if you find owner not trustworthy. So just do it, it’s exactly why it was designed in this way, to let you do it easily.

But talking about funding… I might indeed reconsider doing this…

Is there a reason why you gathered so many features in a single update? It seems to be challenging to test all these features at once, why not just publish them step by step in smaller updates? For instance, Mastodon have even split quote posts into two separate versions (first backend, then frontend) to make this process smoother.

It’s not “manufacturered”, so it’s not conspiracy. But it’s from unreasonable and unacceptable high demand.

Who are investors and where is their money from?

If it’s from investment funds, then I’ve got bad news for all of you, it is your money as well.

I hope this “AI” knows what language is used in songs. Sometimes I like to listen to songs in certain language (usually the one I’m actually learning). The potential use case would be to create a playlist with songs of certain language and genre I like.

As for now, it turns out it’s incredibly hard to do. For example even if you find, say, popular French song and use option to “find similar” then algorithm usually finds either songs with similar genre or even french songs but sang in English.

We evaluated Devstral 2 against DeepSeek V3.2 and Claude Sonnet 4.5 using human evaluations conducted by an independent annotation provider, with tasks scaffolded through Cline. Devstral 2 shows a clear advantage over DeepSeek V3.2, with a 42.8% win rate versus 28.6% loss rate. However, Claude Sonnet 4.5 remains significantly preferred, indicating a gap with closed-source models persists.

Thank you for being honest about performance

Is Multi-Community UI already available? I can see it is merged, but is it released?

and DO NOT DEFINITELY pay for journalist’s work, because he certainly do not need to pay taxes, rent and will happily write articles for you free of charge and paywall is just to make you mad

Stack Exchange. I know, controversial, as people complain about rules being there too strict and community not being too welcoming nowadays, but still a real goldmine of knowledge. All of that with no ads, no spam, no dark patterns.

Even if you go to Europe (where I live), you will still listen to American music, read news about America everywhere, eat American crap food, use American Internet services, watch American TV shows. It is not easy to leave America :D

I wish there was something like this for tags. So: display only posts that contain tag X, but sort them with some algorithm. Ideally: steered by number of likes and date of adding. May be even the same that is used by Lemmy.

… and must share search data with rivals.

What does it mean exactly?

Can I use it? And if not: when can I use it?

State-of-the-art LLM agents do not perform calculations, they call external tools to do that.

To be fair, not all knowledge of LLM comes from training material. The other way is to provide context to instructions.

I can imagine someone someday develops a decent way for LLMs to write down their mistakes in database and some clever way to recall most relevant memories when needed.

I support this request, even issue on GitHub does not contain any (even short) description

+1, I came here just to paste a link to it

Honestly, I thought Meta talking about Fediverse integration was just marketing bullshit. Are they really doing it? 🤔