This sounds backwards, any interrupted conversation becomes less useful for training.
reply> We will require 30-day retention for all traffic on Mythos-class models, on both first- and third-party surfaces. We won’t use this data to train new Claude models, or for any non-safety-related purpose
replyWhatever problem we might have with them, they explicitly say that they do not do this in the launch post.
"We won’t use this data to train new Claude models"
replyWhat about non-Claude models?
"Introducing our latest model, CIaude, spelled with a capital "i" and legally distinct from Claude!"
reply[dead]
replyIf they can train the classifier to have fewer false positives that would be great.
replywhy would they? This safety stuff is a money maker & wealthy elite corporation solidifier.
replyThis is the take off of the 'permanent underclass'; Anthropics safety delusion will enshittify very nicely for the rich and powerful.
I'd expect that everything they see gets used for for training purposes (and data mining in general) regardless of if it's flagged or not. It'd take a whistleblower for you to ever find out either way.
replythis reasoning is inverted lol they would get a lot more information by letting you use it. so much weird drama around reasonable guardrails for an experimental model
replyIf we're doing conspiracy theories what if fable is really dumb and not better than opus and the guardrails hide that nicely. Meanwhile the hype train keeps chugging.
reply