Ornith-1.0: Self-Scaffolding LLMs for Agentic Coding
https://deep-reinforce.com/ornith_1_0.htmlOpen linkView original on lemmy.mlSyndicated from the fediverse. Read and engage on the original instance.
View original on lemmy.mlhttps://deep-reinforce.com/ornith_1_0.htmlOpen linkView original on lemmy.ml
6 replies
It´s is actually impressive: here the paid version of gpt 5.5: and the best it could do (not image generation but coding an svg): then here (down in pic 3) the ornith running locally:
that's pretty good actually
Yes is quite good. I see that they use same techniques like reddit here by downvoting certain posts and shadowbaning to keep them invisible like the idiot downvoting this post. Why someone would downvote a post like this, who may do that! So lame... should we change platform again pfff?
The issue is that It is easy to manipulate the content in this platforms which leads to oppression and missinformation or better forced narrative. That is so easy with just a small bot army or a team of youngsters. Then whats the point of a decentralized platform? Anyway I am thinking to create my own platform instead (may have some better ideas), good suggestion thanks :)
some people just get really upset about anything LLM related I guess
"Pre-famulated amulite for cardinal grammeter automatic synchronisation"