I'm going to use a number of really dumb analogies to try to explain how these large language models like ChatGPT work. First, you have the amount of training they have done combined with the number of things they can look at at once.  ChatGPT 4 is much better than ChatGPT 3 in that regard.  Think of this as the AI's "hard drive". Second, and often unappreciated, is the number of tokens that the model can handle.  Think of this as RAM.  Tokens represents how much the AI...

I'm going to use a number of really dumb analogies to try to explain how these large language models like ChatGPT work. First, you have the amount of training they have done combined with the number of things they can look at at once.  ChatGPT 4 is much better than ChatGPT 3 in that regard.  Think of this as the AI's "hard drive". Second, and often unappreciated, is the number of tokens that the model can handle.  Think of this as RAM.  Tokens represents how much the AI...

I'm going to use a number of really dumb analogies to try to explain how these large language models like ChatGPT work. First, you have the amount of training they have done combined with the number of things they can look at at once.  ChatGPT 4 is much better than ChatGPT 3 in that regard.  Think of this as the AI's "hard drive". Second, and often unappreciated, is the number of tokens that the model can handle.  Think of this as RAM.  Tokens represents how much the AI...

[confluence title=""]https://stardock.atlassian.net/wiki/spaces/SDEnt/pages/1477541924/GalCiv+IV+Supernova+Dev+Journal+10+-+Little+Things[/confluence] ______________________________________________________ [confluence title=""]https://stardock.atlassian.net/wiki/spaces/SUP/pages/1471414278/GalCiv+IV+Supernova+Dev+Journal+Links[/confluence] ______________________________________________________ [confluence title=""]https://stardock.atlassian.net/wiki/spaces/SDEnt/pages/1477541924/...