I'm interested in the position that human beings have not achieved recursive self improvement. I think this may need a little bit of discussion.
This is true for units of the species, but may not be true for societies. This may important for the near term. Given a generation (i) of (n) agents, we can have a multiagent technology (m) st generation (i +1) is better at building multiagent technology (m+1) for the same n agents. The cognition unit here is the society of agents.
Could this be counted as valid RSI? If one were to ban RSI , would your ban cover this research?
I'm interested in the position that human beings have not achieved recursive self improvement. I think this may need a little bit of discussion.
This is true for units of the species, but may not be true for societies. This may important for the near term. Given a generation (i) of (n) agents, we can have a multiagent technology (m) st generation (i +1) is better at building multiagent technology (m+1) for the same n agents. The cognition unit here is the society of agents.
Could this be counted as valid RSI? If one were to ban RSI , would your ban cover this research?