A GPT-4o-mini comparable system that you can run on a RTX 4090 isn’t going to solve direct problems, but it might have enterprise uses. Text generation automation for personal use should be strong, for example - in place of having a third party API do it.

I’ve heard about it before, it’s based on a Skyrim Mod. It isn’t exactly top on my list but I’ll get to it eventually. Thank you for sharing.

You mentioned smaller models achieving better results than ChatGPT, but those models have trouble extending their knowledge to a wide variety of topics, which is shown by their subpar performance in GPQA (general knowledge) tests.

I agree. Coming together to build something scientifical or technological that doesn’t involve the government doesn’t necessarily mean coming up with corporate money to fund and exclude everyone but their personal interests.

This is a social endeavor like any other. You can’t blaim bringing people together for the corporate interests of this kind of enterprise (the Network State).

I think it’s important to come up with other forms of generating synthetic data that doesn’t come from distilling other models. Translating documents, OCRing old documents and using Digital Twins to train visual models come to mind. I’ve never successfully trained any model text-related, but I think the quality of the original text should be critical in how it will perform.

Really liked the article.

Not only the big players extract data from the common citizen, but it also enforces information upon them. AI will make people interact through exchange of knowledge less, and concentrate all the “talk” and information on the hands of few. I think this is a big problem, especially as we near the quantum computation era. How can individuals and smaller organizations possibly compete in AI quality on that scenario? But maybe hardware power won’t be the greatest force in Artificial Intelligence.

It isn’t that bad. It also stated new facts to me, like that GTA is owned by Take-Two.

#ebooks is composed of datahoarders that have a lot of stuff available. You declare the data source you’re getting the book from (e.g. Oatmeal) and then the name of the book.

This is common in rolling releases, but Pop OS isn’t a rolling release distro. Maybe a package you installed or something similar?

Most programming books start with tips for starters. I don’t think that’s a good way to do it. I want books that get into the programming techniques the pros use. What to use when building functions, how to build a proper class within the goals set. I think this is better than explaining the reserved words or other things that belong to an encyclopedia.

Anything that makes Godot better is a good thing.

I also think that reading “A Vampire in my room” with a sexy Chad on the cover isn’t much better than using Instagram, so…

I would blame the Big Tech social networks. My mom uses Instagram all the time and always claims she is getting “a lot of information”. I think that’s to blame. But also reading books isn’t something you force into people, it will make them hate it. It’s just a hobby, either you do it or you don’t. It makes people dumb? Ok. But you can’t force soul-searching (or any other reason to read a book) into someone.

I think that maybe events and other gatherings for readers would be good, people who are searching for something to fill their emptyness but don’t know what yet may get into books (or any other thing).

What a societal mess. Truly something worth of a philosopher’s book.

With a Wi-Fi adapter on Desktop?

Is this possible on Bluesky?

I think OpenAI Whisper is run locally, I’ve used it before. So what’s the catch?

Sometimes LLMs get basic things wrong. This is true.