You are viewing a single comment's thread from:
RE: Are We in an AI Summer or is Winter Coming?
I feel main problems of so called AIs like chat-GPT lie in lack of evaluation of its output. For most people the answers might sound truthfully, but in most cases I went to areas I'm close with I found the presented answers of chat-GPT being vague or even false. So I think as long as chat-GPT or other LLMs not get trained by experts, they will build a fluffy cloud of pseudo informations which people sooner or later get tired of.
Excellent point @ty-ty! Yes, AGI's today are prone to errors and depending on the systems deployed and their purpose, can be very dangerous. However, training these AGIs on human input isn't scalable and hence, I don't see that happening either. In the end, the reason for the next AI winter, may just be cause by AI itself - fascinating stuff! Thanks for commenting.