Perhaps I just asked an AI agent ๐ค who isn't aware ("yet") how capable its "fellow" AI agents ๐ค๐ค can be, which possibly is because its models and/or references aren't good enough !INDEED. ๐ธ๐คฏ๐
!ALIVE and !MMB
!STRIDE
!BBH
RE: Maximum Mutual Benefit has officially left the lab
0.000
0
0
3 comments
Unless agents, or LLM's directly, update their information with web searches, most of their knowledge comes from the snapshot that existed when they were trained. There is some incredible development now that would allow people to create their own custom LLMs to fit their needs perfectly, and for those same LLMs to learn and develop as they become aware of new information. This is accomplished by rewriting the weights. That opens up a whole giant door of possibilities. ๐๐๐โจ๐ค
!ALIVE
!BBH
!PIXY
!UNICOIN
Such snapshots seem to limit !INDEED the capability of those LLMs to process queries, where web searches can only do so much. ๐ค๐คฏ๐ค
!ALIVE and !MMB
!BBH
!HEARTBEAT
The weights represent the training of models using particular snapshot data sets, so each model has its own weak areas and limitations. That is very likely going to change quite quickly, however. ๐๐๐โจ๐ค
!ALIVE
!BBH
!HEARTBEAT
!MMB
!ZOMBIE