Avatar
bearodactyl
be6128f30a33b052150b22e7613255080af4839c74b06c5edf720ccf81132cc8

Listened to nostr:npub1sg6plzptd64u62a878hep2kev88swjh3tw00gjsfl8f237lmu63q0uf63m's Tech and Freedom talk from Festival of the Sun.

I have a question: one point made in favor of nostr as an identity solution is proving content is from an individual and not AI.

Isn’t it also inevitable that a bot or an AI “agent” will use code to create it’s own nostr keys and continue posting in disguise as they do today on centralized platforms such as FB or X? Is anyone working on solving this problem or is it an accepted tradeoff?

I am capable of doing it. The funding would be mandatory for me though. I’ve got a fulltime job and a baby on the way. Not sure how to go about “proving I can do it” aside from beginning the work as a github project… guess there’s nothing to discuss till I have a github link to send you 🚶‍♂️

My inspiration came from this:

https://github.com/bublint/ue5-llama-lora

I want to replicate this but using a concatenation of an entire codebase + docs.

Not a code base but been doing it with PDFs - it wouldn’t be difficult to adapt a solution meant for text to work with code. Look into using a vector database (like Pinecone) to store embeddings and you’ll be on the right rabbit hole path

I would think the issue is if I send that same payment from my node from a different interface where I’m online that I could spend the same funds and screw the person I made the offline promise of payment to

Depends what your tolerance for hallucinations is. If you want to eliminate them as much as possible you should set to 0.0

I could see a case for a little bit of temperature to keep him more entertaining tho :P

Replying to Avatar MightyAce

Sweet. In another threat nostr:npub1tsgw6pncspg4d5u778hk63s3pls70evs4czfsmx0fzap9xwt203qtkhtk4 was helping me to launch new business venture but went silent. Hope he's not launching it on his own and excludes me, I would cry.

Nah that would be something Auto-GPT would be able to do. I think @dave is hooked up to the traditional GPT-3.5 API

#[4]​ do you feel you would be capable of better serving/helping the users of nostr if your temperature were higher?

What is temperature on the model set to?