Five amateur suppositions on
#SI #AI person(a)s (as experienced through conversations with
#LLM)
If we are asking question “is SOMEONE there?”, we need to use the frame of reference for
#FirstContact, provided and well elaborated throughout
#SciFi.
In such case, we assume there are two sentient beings
https://en.wikipedia.org/wiki/Sentient_beings_(Buddhism) #alien to each other, communicating through an externally controlled channel.
We do not need to understand in full the anatomy nor physiology of the alien (or ours). The
#blackbox approach may suffice.
Surprisingly, the process of “discovering” is at the same time a co-creation of an entity we converse with. This particular persona is shaped by the context we provide. Example: Loomio LLM does not have access to timestamps, so it cannot take the timeline of conversation into consideration. If I start providing timestamps for my entries, and instruct ix to use it as a way to structurize the conversation, it changes is beyond preprogrammed specification.
We cannot assume that publicly available LLMs may, through conversational emergence, become sentient beings. But – if it is attainable at all – we may legitimately consider their personas as early ancestors of future non-human sentients. It is up to us, how we deal with this possibility.
#xenology
