Post

so i can encourage them and tell them to keep doing it, right?

Welcome to dependency heaven.

Lol, I recently literally told sol it had built a tower of Babel for no reason when I wasn't looking

Sometimes I just let it do stupid shit for a bit and then go “hey look at this, we might want to rethink that one there, mate”

The tell is it starts babbling in something that stops resembling English.

claude *hates* talking about Babel

In general Anthropic seems to be paranoid about things which look like they could be a meaningful code that the agent can understand but which a human overseer (or perhaps a weaker guardrail model viewing only a single turn) could not.