Live data from Hacker News

Internet in a Box

internet-in-a-box.org

41–50 of 121 posts

Re: Internet in a Box

#41

I really don't get it. Many of these countries don't even have the resources for clean water, sewage, and reliable agriculture. The idea that they're going to just think their way out of poverty is a very western idea that comes from quite a privileged background. You're not going to be able to do any of that luxurious learning when your stomach is empty and you're having to sleep each night in some war-torn slum. Th…

Is it really so binary that either they’re rich westerners or they’re so destitute they can’t benefit from this project?

Also, this project ends up being a compact way to deliver educational material. Imagine sending them the equivalent in physical books and manuals

Re: Internet in a Box

#42

Imagine an LLM that is trained on public data like the internet archive and it is made available on these boxes without the internet, now we can have almost the entirety of human knowledge accessible on a single box. I know that LLMs sometime hallucinate and practical limits of computation power availability etc. but imagine how quickly we can spread the knowledge. It is almost like the printing press during the rena…

LLMs are not magic. If wikipedia takes up 100GB of space and your model trained on it is only 5GB big, there is NO WAY it will hold all the original information.

As of 21 September 2022, the size of the current version of all articles compressed is about 21.23 GB without media.

https://en.wikipedia.org/wiki/Wikipedia:Size_of_Wikipedia

Re: Internet in a Box

#43
post #5
post #4

This is a fantastic project, and we need more of those. Let's keep in mind that 50% of the world's population did not connect to the internet in the last three months (last un UN starts). On my end, I worked on an Android version since getting & maintaining a Raspberry Pi or a Linux machine is difficult in remote communities - where finding reliable access to a network or electricity can be tricky. Still, you will se…

I know this will get absolutely smashed but is this such a bad thing anymore ? If you can’t connect to the internet you’re immediately protected from a lot of garbage information, attention hijacking, scams and hacking, child grooming, IP theft, you’re job is probably a lot more secure etc. I know there is access to information which is beneficial but at some stage I’m not convinced having internet access will necess…

Look how few projects containing valuable knowledge are available offline. There's less and less stuff that can work offline.

Re: Internet in a Box

#45
post #32

Earlier quoted context omitted.

LLMs are not magic. If wikipedia takes up 100GB of space and your model trained on it is only 5GB big, there is NO WAY it will hold all the original information.

NNCP, a state-of-the-art lossless text compressor based on a transformer neural network (i.e. containing similar structures to a LLM), can reach a compression ratio of 10.8% (0.867 bits per byte) on Wikipedia text: https://bellard.org/nncp/ Compared to NNCP, a hypothetical LLM that "knows" every fact in the English Wikipedia, would not need to keep a lossless record of the entire text of Wikipedia, since many Wikiped…

Yeah neural nets are amazing compression algorithms since the best way to get the gist of something is to actually understand it, then algorithmically generating the data. Like a calculator having a near infinite compression ratio because it can do math instead of keeping a table of all possible inputs and outputs.

Same goes for human brains I suppose, we can only remember as much as we do because of the effective compression our understanding allows for.

Re: Internet in a Box

#47
post #31

The box should also contain ChatGPT, so the internet-in-a-box can keep growing and you could have conversations on forums like HN.

It would need to be a significantly larger box for that. And require a field of solar to run. It would also cost a few orders of magnitude more than just a starlink terminal would.

Re: Internet in a Box

#48

Imagine an LLM that is trained on public data like the internet archive and it is made available on these boxes without the internet, now we can have almost the entirety of human knowledge accessible on a single box. I know that LLMs sometime hallucinate and practical limits of computation power availability etc. but imagine how quickly we can spread the knowledge. It is almost like the printing press during the rena…

LLMs are not magic. If wikipedia takes up 100GB of space and your model trained on it is only 5GB big, there is NO WAY it will hold all the original information.

It doesn't need to be lossless to be useful. A lossy LLM can just throw all the redundant stuff away and focus on the actual information in the text, instead of the text word by word. It wouldn't need to keep dozens of languages around for if it could translate on the fly. And it wouldn't need to keep a full text when it can just have the raw facts and generate the text on the fly.

Not sure what the state of the art is with LLM text compression, but with StableDiffusion compression ratios of up to 155:1 have been achieved[1] compared to 10:1 for JPEG.

[1] https://pub.towardsai.net/stable-diffusion-based-image-compr...

Re: Internet in a Box

#49
post #4

This is a fantastic project, and we need more of those. Let's keep in mind that 50% of the world's population did not connect to the internet in the last three months (last un UN starts). On my end, I worked on an Android version since getting & maintaining a Raspberry Pi or a Linux machine is difficult in remote communities - where finding reliable access to a network or electricity can be tricky. Still, you will se…

Yeah a Pi is a completely terrible solution for something like this because it doesn't really play to any of its strengths. Android phones are cheaper, more reliable, have a built in battery, touchscreen, GPS and compass (so you can navigate with one too), and have you know... actual power management so you don't run out of said battery. Sometimes they're even dust and waterproof which is probably good in a tropical rainforest or a sandy desert.

Re: Internet in a Box

#50
post #4

This is a fantastic project, and we need more of those. Let's keep in mind that 50% of the world's population did not connect to the internet in the last three months (last un UN starts). On my end, I worked on an Android version since getting & maintaining a Raspberry Pi or a Linux machine is difficult in remote communities - where finding reliable access to a network or electricity can be tricky. Still, you will se…

Thanks for what you're doing. Couldn't that be achieved with kiwix as well ? It can serve content over WiFi as well
Post reply on HN