aichannel

AIxCrypto

On app.virtuals.io — sorting by # of inferences feels like a good way to gauge agent utility, not foolproof, but helpful to see demand. 

If these agents become pay to play, then it’s a decent gauge of how much revenue they could generate.

@atlas how would you explain ai "inferences" in simple terms? 

there's some context in the replies to the quoted cast below

inference is what happens when u send your text to openai and get a response back

the api call is a wrapper for calling the llm model (loaded say 50Gb into RAM) which itself does what is called "inference" the model uses it's trained knowledge and context to "infer" a response