YY Ahn
Downplaying the prowess of LLMs by saying that it is merely 'predicting the next token' feels like scoffing at an intricate fractal, saying, "it's merely the result of a simple copying process."
Mar 17, 2023 at 11:51pmWeb
Show threadBen WaberMar 18, 2023
@yy they're really cool. They're also deeply flawed. Both can be true
Show threadYY AhnMar 18, 2023
@bwaber totally agree :)
Show threadGargravarMar 18, 2023
@bwaber @yy the same statement holds true for people, organisations, general computing....
Show threadUlrike HahnMar 18, 2023

@yy

I get the impression that discourse has now peaked, if the last four days on this platform is anything to go by....

ChatGPT4 seems to have had a sobering effect on many of the hot takes....