Post
Joe Barnby
joebarnby.com
did:plc:jgawqxiexrmiaaa4wjssj5pj
What happens when we can't use recursive belief to compete? We can use anomaly detection instead!
Here, we (led by soon-to-be-Dr @nitalon.bsky.social ) devise a multi-agent account where compression & reward expectation are used to notice deception
https://jair.org/index.php/jair/article/view/19204/27269
2026-02-18T08:49:15.580Z