Help. That’s it, that’s all I’d say

Nah, it's 22 GB for just the current English version, text-only Wikipedia. Adding the media English current version only would be 140 GB (plus the 22 GB for the text). https://en.wikipedia.org/wiki/Special:MediaStatistics https://en.wikipedia.org/wiki/Wikipedia:Size_of_Wikipedia

So you could do text-only Wikipedia and probably compress it. Maybe drop a few thousand of the articles that don't matter or are stubs. Drop all the entertainment articles, etc.

I don't think you appreciate how much text you can fit into 4GB. The first entire gigabyte could be dedicated to various means of translation and explaining our language system, and you'd still have a 500 million words left after that.

Can agree with that, but I wanted to save the alien from having a Microsoft account.

No reason at all. Of course probably the aliens won't have access to the interwebz, but playing Minecraft with an alien sounded funny in my head.

It's explained in the PDF. If they used normal numbers and letters, the chance of corruption would be high. So they had to reform into the symbols you see.

I still don't get why we use Pi instead of Tau though, when most equations double it up into Tau anyway.

So that the symbols would be less likely to get corrupted if the aliens received it over a bad radio link

Because an LLM is more than just data: it's like a big network of how syllables and words go together based on some context. And that's useful because language is how we communicate, how we connect ideas together, it's how we share stories. It's not just Wikipedia articles, it's a database of relationships between words and concepts. It approximates how we think as humans.

Yes, AI is hella overhyped. Everyone wants to AI everything. But really for this particular situation, I think the model data would actually be the best precompiled database of knowledge we can possibly provide to learn about humans for the size.

No it's not magic compression, but 4GB worth of parameters is still a lot. GPT4All has models just under 4GB. They're not particularly impressive compared to OpenAI's offerings, but I think you can extract a lot more practical information to do first contact out of a basic model than 4GB worth of Wikipedia. It's extremely lossy compression, it's never gonna spit out articles vebatim, it will hallucinate a ton of stuff.

If we had more space I'd send all the major AIs we have like Dall-E, LLaMa and GPT 4. Imagine you're an alien, you're presented with a keyboard and a monitor, and know nothing about us. You can use Dall-E to try random letters and words and see if the output makes sense. Maybe you find out what a cat, dog, bat, frog, apple looks like. You can then input those words in ChatGPT, and get context as to when those are used. What's "a horse"? What's "riding"? Put those into Dall-E, now you know what a "cat riding a horse" looks like. It can generate as many as you want, any combination. Eventually you can figure out how to ask ChatGPT if cats typically ride horses, cars, bycles, what do cats do.

Now imagine you're a very advanced alien species that can easily process the model's parameters. You've just downloaded the basics of humanity. They can map their language to our model's parameters, and basically speak to us in our language, and translate our answers to theirs, and basically have a basic conversation.