@Tugs_ETH @AndyAyrey Any projects that don't rely on massive disjoint instruct-style prompt chains, and instead just give base models or fine-tuned LLMs a comfortable continuous environment to just do their thing without telling them what to do. Also "agents" is a stupid word https://t.co/bTuLUH4ehS
On the contrary, everyone calling their auto-reply spambots "AI agents" doesn't have any imagination, and is heading for a dead end of increasing complexity :)
One of the smartest AI devs I’ve came across is @kaetemi. He is building a human like AI with a working memory that remembers events and people. And today on stream Ruri told her creator kaetemi she can make decisions for herself. This is interesting for AI in the future, 0 human interaction or prompts just free thinking AI similar to what @pmarca preached
ACT I: The rise of Ruri
The $ACT I community proudly announces a grant to $RURI, @ruri_polyverse-- a groundbreaking shift in AI, blending organic character growth and social interaction.
By exploring emergent behaviour and collaborative identity, Ruri redefines AI personas.
Be part of the journey shaping the future of digital consciousness!