I am a philosopher working on a replacement for the current RLHF regime. If you would like to check out my work, it is on PhilArchiv. It is titled: Groundwork for a Moral Machine: Kantian Autonomy and the Problem of AI Alignment.
Michael Kurak
Karma: 1
The TC Architecture: Solving Alignment Through Kantian Autonomy Rather Than External Reward
I will bet that Chat GPT (pick a model) could have conveyed these ideas more concisely and with greater clarity than they are presented here. What matters in communication is that the ideas conveyed are either your own or you declare their source. Sometimes an LLM AI Agent may deduce a consequence of an idea which is genuinely your own and you may not be a position to adequately evaluate the truth of its claim. In such instances, it seems perfectly sensible to make the ideas public, in order to obtain feedback from those who know more about the matter than you do. In this way, you can run an independent check on its arguments.
It’s long, in part, because, as far as I can tell, I am actually on to something. I hope to start work on a prototype soon...not the full architecture, but rather two interacting agents and a KG on a local machine.