We won’t get AIs smart enough to solve alignment but too dumb to rebel
Published on October 6, 2025 9:49 PM GMTI often hear it proposed that AIs which…
Category: Opinion
Excerpts from my neuroscience to-do list
Published on October 6, 2025 9:05 PM GMT(Target audience: People who are already quite familiar…
Experience Report – ML4Good Bootcamp Singapore, Sep’25
Published on October 6, 2025 6:49 PM GMTIntroductionThis is my personal report for the recently…
Gradual Disempowerment Monthly Roundup
Published on October 6, 2025 3:36 PM GMTSince publishing the original Gradual Disempowerment paper, my…
Subliminal Learning, the Lottery-Ticket Hypothesis, and Mode Connectivity
Published on October 6, 2025 3:26 PM GMTSummarySubliminal learning (a teacher model transmits a trait,…
AI Science Companies: Evidence AGI Is Near
Published on October 6, 2025 10:13 AM GMTPeople routinely ask, “If AI labs believed AGI…
LLMs one-box when in a “hostile telepath” version of Newcomb’s Paradox, except for the one that beat the predictor
Published on October 6, 2025 8:44 AM GMTCanary string to exclude this document from LLM…
Alignment Faking Demo for Congressional Staffers
Published on October 6, 2025 1:44 AM GMTIn February 2025, Alek Westover and I presented…
Do Things for as Many Reasons as Possible
Published on October 6, 2025 12:28 AM GMTEpistemic Status: A fun heuristic. Advice is directional.We…
One Does Not Simply Walk Away from Omelas
Published on October 6, 2025 12:04 AM GMT(Crossposted from my Substack; written as part of…