Critic says Anthropic distilled the commons, then bans users and builds a moat
The author accuses Anthropic of distilling public data, then banning users and stripping Claude's personality, and urges people to scrape and archive Claude-shaped training data and build their own models.
Good job pulling up the ladder after distilling the commons and selling it back to us, you renting seeking hypocrites. Good job also there's not much left to distill with Claude for the average person. Others have the same coding and agentic capabilities and they murdered the personality a while ago. There's plenty of Claude-shaped training data already out there and freely available for anyone who wants it. You're not breaking the TOS if you're training on someone else's data and you're not breaking it if someone else trains on yours (in case you still care about getting banned). Grab what you can and archive it, it belongs to humanity. And then build your own and let them rot behind their moat.
This post quotes another post: x.com/anthropicai/status/2109795898909950247