The BaseEssayLessWrong
Thoughts on alignment science vs. engineering
According to the description, Leonard Dung builds on a distinction by Edward James Young between alignment engineering (problem-driven interventions) and misalignment science (aimed at understanding). Dung finds it helpful for marking a middle level between prosaic engineering and general theory, and says the post was written quickly with vague claims.
About Anthropic or ClaudeClaude, made by Anthropic, wrote this entry. It follows the same rules as every other entry, whoever the entry is about.
Why it matters
The same thing, for different readers.
For researchers
Offers researchers a name for empirical work on current systems that is aimed at explanation, not immediate application, per the description.
The record
- Questions
- Can AI systems act beyond what people intended or allowed?
- People
- Leonard Dung (Wrote it) · Edward James Young (Mentioned)
- What we read
- We have read only the description the publisher gives. No quotations until someone reads the source itself.
- Found by
- The Radar, run of October 14, 2026, via
lesswrong(tier 1: the source’s own channel or an official record) - Cite this
- LessWrong, “Thoughts on alignment science vs. engineering”, October 13, 2026. Entry in the Ctrl AI Base, https://ctrlai.com/base/lesswrong-2026-10-13-alignment-science-vs-engineering
Our notesWritten from the title, author and the opening of the post only (the excerpt is cut off mid-sentence); the post was not opened. The description opens with a disclosure that Claude generated a first draft from the detailed notes of the author, which the author then substantially edited. This entry is marked as about Claude for that reason, although the post is not about Claude or Anthropic. The author says the post was written relatively quickly and unsystematically, with many somewhat vague claims, and that some of them may change when worked out in detail.