That claim about load bearing is doing a lot of work there!
I'm curious to know how "Claudish" emerged during the training process. Why each AI has a particular voice if much of the training material is the same across AIs?
They train Claude on its own output and conversations with users, so tics like this are going to get amplified by virtue of being more common in the training set.
That, and my conspiracy hat says it's a way of watermarking the output.
Comments
That claim about load bearing is doing a lot of work there!
I'm curious to know how "Claudish" emerged during the training process. Why each AI has a particular voice if much of the training material is the same across AIs?
They train Claude on its own output and conversations with users, so tics like this are going to get amplified by virtue of being more common in the training set.
That, and my conspiracy hat says it's a way of watermarking the output.
NGL i also suspect the particular brand of Anthropic safety filters cut out passionate/fun personality traits