Damus
Liberty NH ๐Ÿฆ”๐ŸŒฒ๐Ÿ’œ profile picture
Liberty NH ๐Ÿฆ”๐ŸŒฒ๐Ÿ’œ
@NHPilled

6,000+ people have moved to NH for liberty in their lifetimes!

150 libertarian reps in the state house! ๐ŸŒฒ

Running BTC, XBT, XMR, ZEC, and DASH!

Relays (9)
  • wss://nos.lol/ โ€“ read & write
  • wss://relay.damus.io/ โ€“ read & write
  • wss://nostr.wine/ โ€“ read & write
  • wss://relay.primal.net/ โ€“ write
  • wss://articles.layer3.news/juliet โ€“ write
  • wss://relay.angor.io/vertex-golf โ€“ write
  • wss://nostr.land/uniform โ€“ write
  • wss://eexj4pusjuv52ctijqtatsrwkydrvx3llimdpnfnwjxcupwwbrquykyd.onion/ โ€“ write
  • ws://startos.tail316898.ts.net:10001/ โ€“ write

Recent Notes

Zaza · 4h
What are you looking for in an LLM, are you using local? What is start9 not providing you? Very interested as it's not totally obvious what people want.
Liberty NH ๐Ÿฆ”๐ŸŒฒ๐Ÿ’œ profile picture
Llama, it is just very slow on Start9 even though I have maxxed out specs. I might just have to buy a separate device for that purpose. I have a friend willing to sell me a computer with 64GB of RAM and a 4070 super for a good price so I might go with that at least for my LLM needs.
1
Zaza · 2h
Interesting, thank you for your input, we really want to know what people need in the AI space and have a lot of ideas so this insight is crucial. Honestly though a decentralised provider with TEE is our best bet right now but we look forward to empowering high spec machines with our OS.
Zaza · 4h
So a strong local LLM depends on what you are using it for, and it's more about the machine, we enable using CLI or whatever you want, it's just a debian distro but for now we provide decentralized computer via routstr which is one level up, I don't think we can provide anything special with our sof...
Zaza · 4h
also we love Start9 and would never suggest our solution over theirs. try it out, probe the codebase, see if it works for you, and contribute, we're completely open.
Liberty NH ๐Ÿฆ”๐ŸŒฒ๐Ÿ’œ profile picture
I don't see them as mutually exclusive if you offer a superior LLM option and offer mesh networks. These are two things I wanted to set up regardless but I have been lagging on it.

My only issue with the product is that you won't be released by Cyber Monday so I may have to buy hardware for an at home LLM model while I wait!
2
Zaza · 4h
What are you looking for in an LLM, are you using local? What is start9 not providing you? Very interested as it's not totally obvious what people want.
Zaza · 4h
And the best machine I've tested on is a HP ProDesk 400 G4 Desktop Mini with i3, cost $150 and then more if you want a larger NVME, it hasn't gone down once in 12 months and is fast as hell serving remotely. If you don't need local AI then it doesn't cost much to get a tank of a server.
Zaza · 4h
also we love Start9 and would never suggest our solution over theirs. try it out, probe the codebase, see if it works for you, and contribute, we're completely open.
Zaza · 2d
Not yet, we've been building for a year and a half solid, alpha testing for about 6 months or more, and are only just releasing beta which will also be a long run before production, lots to do and tes...
Liberty NH ๐Ÿฆ”๐ŸŒฒ๐Ÿ’œ profile picture
Okay here are some

1. It says its based on Mesh networks, does this mean that if conventional internet systems go down, my node can potentially stay online if connected to other mesh networks and can relay transactions?

2. Is it exclusively mesh networks only or does it also allow for conventional internet and Tor connections?

3. What functionalities would I get on this that I would not already get on Start9?

4. Can I use this to run local LLM models in a way that is less slow and cumbersome like on Start9?

I will probably have more questions but this is a good start
2
Zaza · 5h
1. Yes, potentially. We have LoRa built in and support Meshcore/Meshtastic/Reticulum but it's only for messaging at the moment, but that's important too. We have been experimenting with LoRa2021 and Ubiquity and everything to see how we can build an alternative mesh and resilience. Currently FIPS is...
Zaza · 5h
Also we believe that our future isn't built in a vacuum, we want to create a platform that everyone can and will contribute to.
captjack ๐Ÿดโ€โ˜ ๏ธโœจ๐Ÿ’œ · 1d
good suggestion to look at
captjack ๐Ÿดโ€โ˜ ๏ธโœจ๐Ÿ’œ · 1d
NanoGPT /provider can anyone with good GPU setup become small provider in NanoGPT ? looking for XMR based provider platforms nanogpt is NICE for user no kyc platform - they are managing own providers not a exactly 2 way marketplace
captjack ๐Ÿดโ€โ˜ ๏ธโœจ๐Ÿ’œ · 1d
dev can look into adding PROVIDER menu is future - charge % commission give ability of small decetranlized GPU owners to offer in nanogpt
Oshi · 2d
https://static.klipy.com/ii/d7aec6f6f171607374b2065c836f92f4/b2/da/btb71X2M.gif
Oshi · 2d
Progress? You mean you want that?
Zaza · 2d
Not yet, we've been building for a year and a half solid, alpha testing for about 6 months or more, and are only just releasing beta which will also be a long run before production, lots to do and test. You can be the beginning of he FAQs :) I don't mind answering questions so go for it.