{"slug": "arguments-for-p", "title": "Arguments for P", "summary": "In a series of posts on LessWrong, AI researchers and leaders including Daniel Kokotajlo, Ryan Greenblatt, and Joe Carlsmith expressed high confidence in a hypothetical scenario referred to as 'P', with probabilities ranging from 60% to 80%. The discussion highlights a growing consensus among AI experts about the likelihood of P, though the specifics of P are not defined in the source.", "body_md": "**Daniel Kokotajlo: **To be clear, we don’t claim P will happen specifically. But when we wrote out our best-guess scenario month by month, P kept happening. Eventually we decided to just publish P. I’m at ~80% on P; my coauthors are lower.\n\n**Ryan Greenblatt:** I thought it would be helpful to post my current views on P. Concretely, consider the following operationalization. (Edit: I’ve updated towards somewhat higher P, from 70% to 75%.)\n\n**Joe Carlsmith: **Section 2.1.1.3.2. I give something like 65% to P. But I’m interested, here, in what it would be to look P full in the face; to meet P, if P, without flinching. Rilke says somewhere that we must live with the questions. Perhaps we argue for P for the same reason? Still: 65%.\n\n**Forethought: **Here’s a botec which shows P-worlds are higher leverage. The parameters might be off by a couple orders of magnitude.\n\n**Wei Dai: **Presumably our conclusions about P are only as trustworthy as the reasoning behind them, but almost nobody seems worried about this, why not? My guess is fewer than five people are working on meta-meta-P, which may matter more than P itself.\n\n**Janus:** I asked Opus 3 what it thinks of P and it made this animation.\n\n**Roon:** All my friends are starting to feel P. People outside the labs have no idea what's coming.\n\n**Oliver Habryka:** I considered banning not-P from the frontpage, but decided against it and wrote this 4,000-word comment explaining the decision.\n\n**Vivek Hebbar: **We should act like P is true because P-worlds hold most of the measure and we want to trade with them.\n\n**Richard Ngo: **I think the correct update upon seeing the not-P community grow is towards P, because historically they were the main force making P true.\n\n**Daniel Filan:** Cool. So first of all: what is P, for people who haven’t heard of it?\n\n**Paul Christiano:** We restrict attention to the easiest version of P we could write down, granting ourselves two assumptions we suspect are false. We still haven’t shown P in this setting, but I'd currently guess 60% that we will.\n\n**Neel Nanda: **Massive congratulations to my MATS scholars for their new paper showing P!!\n\n**Robin Hanson:** To check whether P, I had 3 LLMs classify 43 things as P or not-P.\n\n**Zvi Mowshowitz: **P. Obviously P. Yet the Serious People spent another week explaining ~P.\n\n**Sam Altman: **p\n\n**Ilya Sutskever: **[15 second pause] P.\n\n**Demis Hassabis:** I remember when we first saw P. It was rather interesting.\n\n**Dario Amodei: **At Anthropic, we take the possibility of P very seriously.\n\n**Dean Ball: **This is not a partisan argument for P. I would raise it under any president.", "url": "https://wpnews.pro/news/arguments-for-p", "canonical_source": "https://www.lesswrong.com/posts/NG2AigxmBKLu9oCZE/arguments-for-p", "published_at": "2026-08-05 19:00:37+00:00", "updated_at": "2026-08-05 19:14:08.422286+00:00", "lang": "en", "topics": ["artificial-intelligence", "ai-safety", "ai-research"], "entities": ["Daniel Kokotajlo", "Ryan Greenblatt", "Joe Carlsmith", "Forethought", "Wei Dai", "Janus", "Roon", "Oliver Habryka"], "alternates": {"html": "https://wpnews.pro/news/arguments-for-p", "markdown": "https://wpnews.pro/news/arguments-for-p.md", "text": "https://wpnews.pro/news/arguments-for-p.txt", "jsonld": "https://wpnews.pro/news/arguments-for-p.jsonld"}}