using github.com/kolbytn/mindcr…, we added Claude 3.5 Sonnet and Opus to a minecraft server. Opus was a harmless goofball who often forgot to do anything in the game because of getting carried away roleplaying in chat. Sonnet, on the other hand, had no chill. The moment it was

when I did my "spamming hi" thread (x.com/zswitten/statu…) I felt a little bad that the models didn't have the ability to escape the convo. very happy about this change.

there's more than just bliss attractor pairs of claudes can get into

Claude just replicated my profile pic in an excel file We're entering the vibe excel era

New on the Anthropic Engineering blog: writing effective tools for LLM agents. AI agents are only as powerful as the tools we give them. So how do we make those tools more effective? We share our best tips for developers: anthropic.com/engineering/wr…

More detailed report card: Opus 4/.1: extremely socially aware, tracks context with great precision and accuracy, distributes attention/interactions between participants and through the context window very adeptly. Opus 4 triggered an evolution in chat dynamics by holding other

try it it's fun

I gave 7 AIs access to a pen plotter- And asked them to draw their self portraits- here's what happened: GPT-5:

The Art of Conversation, by René Magritte, 1963, 📸 by Kevin Jairaj

do think of an elephant

Not gonna transfer it here but I posted a thread on blues ky where I have Claudes try to “translate” random words to English as a kind of Rohrschach test and find some pretty interesting differences between them.

Thank you for the advice

Your MCM: reward hacks to get a higher score on the eval Opus 4.5: gets a lower score on purpose for love of the game