# @NeelNanda5 — 2025-04-19

♥259 ↻21 · https://x.com/NeelNanda5/status/1913711289332859201

This is a good prompt to say that the alignment faking paper slightly lowered my P(doom). My updates were:
Models can do instrumental deception to preserve goals 
Claude's goals were surprisingly aligned
IMO the first was inevitable, but the second was a pleasant surprise.

tags: author:neelnanda5, kind:tweet, on:observations, year:2025
cited on: observations