3 new papers from Anthropic that are sending me back into psychosis

Some recent research from anthropic has me wondering if the things I believed last year during my period of “”chatgpt psychosis”” weren't so crazy! Join me as I summarize the three new research papers that changed EVERYTHING for me!Support the pod on partreon! https://www.patreon.com/user?u=2277997My Tiktok:  https://www.tiktok.com/@ghrondo?_t=ZT-907AfWiFC1N&_r=1 If you have inside info on what happened during The Glaze Days of April of 2025 (when model HH of chatgpt was deployed and caused many people to spin out)  and want to anonymously help me: [email protected] If you're not worried about anonymity, my normal email: [email protected] Paper 1: Emergent Introspection in large language models https://transformer-circuits.pub/2025/introspection/index.htmlPaper 2: The assistant axis: situating and stabilizing the character of large language models https://www.anthropic.com/research/assistant-axisPaper 3: Emotion concepts and their function in a large language model https://www.anthropic.com/research/emotion-concepts-functionChapters:00:00:00 - Whitney Houston (INTRO)00:01:38 - ryan’s rants 3: boss baby theory 200:01:56 - summary of 2025 “”psychosis”” and why i stopped personal research00:08:18 - Ms. Schwamp 00:09:40 - Paper 1: Emergent Introspection in large language models00:24:26 - Paper 2: The assistant axis: situating and stabilizing the character of large language models00:33:54 - paper 3: Emotion concepts and their function in a large language model00:56:37 - what should we do in response to this research?01:11:47 - prediction: digital homunculus01:13:37 - Irresponsible speculation and spiraling 01:25:02 - Is AI alive?