someone on Nostr: using unsloth to pretrain with Nostr notes. https://unsloth.ai/blog/contpretraining ...
using unsloth to pretrain with Nostr notes.
https://unsloth.ai/blog/contpretraining
my code is similar to what they shared in the Colab notebook.
this will be one of the most based models on the planet, powered by nostriches.
with more and more based models we will define what is beneficial to humanity. this will allow us to align AI with humanity better than other alignment projects.
thanks to Nostr's ability to attract the truth, now AI just has to memorize the notes in order to stay aligned. with proper alignment we are looking at an AI not just beneficial, but could also protect humans against a probable evil AI.
https://unsloth.ai/blog/contpretraining
my code is similar to what they shared in the Colab notebook.
this will be one of the most based models on the planet, powered by nostriches.
with more and more based models we will define what is beneficial to humanity. this will allow us to align AI with humanity better than other alignment projects.
thanks to Nostr's ability to attract the truth, now AI just has to memorize the notes in order to stay aligned. with proper alignment we are looking at an AI not just beneficial, but could also protect humans against a probable evil AI.