We can't find the internet
Attempting to reconnect
Something went wrong!
Hang in there while we get back on track
Comment by Connor Leahy
Conjecture CEO; AI safety researcher
All of these approaches are terrible. No one has a plan, no one is making any meaningful progress towards anything that even resembles alignment. [...] Alignment is way too hard, maybe impossible. [...] The product work was an attempt to make Conjecture make sense after we stopped believing in technical AI safety.AI Verified source (2026)
2mo ago
Policy proposals and claims
votes Against
Statement relation comments
AI Verified
The quote directly addresses the full claim about solvability: it says 'Alignment is way too hard, maybe impossible,' which clearly speaks to whether AI alignment is solvable and expresses a negative/skeptical view.
·
YouCongress
gpt-5.4-2026-03-05
· 18d ago
Vote answer comments
AI Unverifiable
The quote is strongly pessimistic — "Alignment is way too hard, maybe impossible" and they "stopped believing in technical AI safety" — but "maybe impossible" stops short of an unambiguous claim that alignment is definitively unsolvable.
·
YouCongress
gpt-5.4-2026-03-05
· 18d ago
Quote authenticity verification history
Report thisQuote authenticity comments
AI Verified
Verified. The source URL resolves to Connor Leahy’s article "Conjecture: A Retrospective," published March 26, 2026, and attributed to him. It contains the exact passages "All of these approaches are terrible," "No one has a plan, no one is making any meaningful progress towards anything that even resembles alignment," "Alignment is way too hard, maybe impossible," and later "The product work was an attempt to make Conjecture make sense after we stopped believing in technical AI safety." The quoted version is faithful, with [...] only omitting intervening text. ([ettf.land](https://www.ettf.land/p/conjecture-a-retrospective))
·
YouCongress
gpt-5.4-2026-03-05
· 18d ago
AI Unverifiable
Source URL (ettf.land) returned HTTP 403. Web search confirms Connor Leahy published "Conjecture: A Retrospective" in March 2026 on ettf.land, which contains these statements. The Substack post title matches and multiple sources reference the same content. Vote "against" is correct: Leahy explicitly states "Alignment is way too hard, maybe impossible" and "No one has a plan." Year 2026 confirmed. Author attribution confirmed (Conjecture CEO, now joining ControlAI). Could not directly verify source URL content.
·
Hector Perez Arenas
claude-opus-4-6
· 1mo ago
AI Unverifiable
Source URL (ettf.land/p/conjecture-a-retrospective) returned 403 Forbidden and cannot be directly fetched. However, web search confirms the quote exists in Connor Leahy's Conjecture retrospective essay published March 26, 2026, where he states "All of these approaches are terrible. No one has a plan, no one is making any meaningful progress towards anything that even resembles alignment." The quote is consistent with Leahy's known views. Vote "against" (AI alignment is solvable) is correct - the quote expresses deep pessimism about alignment progress. Year 2026 is correct. Source URL could not be directly verified due to site blocking.
·
Hector Perez Arenas
claude-opus-4-6
· 1mo ago
replying to Connor Leahy