3 ms·
These guardrails are solely a reason for using your data for training purposes. Every flagged message can be used for training.
by I_am_tiberius 5mo ago
These guardrails are solely a reason for using your data for training purposes. Every flagged message can be used for training.
- wmf 5mo agoIf they can train the classifier to have fewer false positives that would be great.
- cyanydeez 5mo agowhy would they? This safety stuff is a money maker & wealthy elite corporation solidifier. This is the take off of the 'permanent underclass'; Anthropics safety delusion will enshittify very nicely for the rich and powerful.
- Retr0id 5mo agoThis sounds backwards, any interrupted conversation becomes less useful for training.
- autoexec 5mo agoI'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.
- make3 5mo agothis 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
- tekacs 5mo ago> 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 Whatever problem we might have with them, they explicitly say that they do not do this in the launch post.
- Merik 5mo ago"We won’t use this data to train new Claude models" What about non-Claude models?
- flexagoon 5mo ago"Introducing our latest model, CIaude, spelled with a capital "i" and legally distinct from Claude!"
- MagicMoonlight 5mo ago[dead]
- Lord_Zero 5mo agoIf 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.