Prime target: reaching AGI with the brain inspired architecture.
The approach seems logical when understanding that the LLM’s are to large part based on past text written by humans or robots (AI) defining the ‘status quo’ which reflects rather a messy world. The human native brain on the other side is the most sophisticated existing thinking entity. In last decades, via internet, knowledge was shared between billions of people (brains) resulting in accelerated science progress and tech revolution. Hence, Sapient approach, despite also using LLM portion is building on the strongest (verge) base ingredient, the human brain. Human brain performance is dependent on biological, evolutional and long learning (study/school period) limitations. Sapient solution, in case AGI is reached, may scale much faster then human brains (accelerated evolution). This scaling speed represents the risk for humans in particular with the Sapient model.
AGI Scaling guidelines are to be established to ensure the evolution leads to a human friendly future and does not morph into a total mess for humans where they won’t understand what is happening.