HackerPapers

arXiv:2603.15973cs.AI

Safety is Non-Compositional: A Formal Framework for Capability-Based AI Systems

Cosimo Spera

Early discovery. Linked by one identifiable commenter. This paper is outside the main ranking; a mention is not an endorsement.

Abstract

This paper contains the first formal proof that safety is non-compositional in the presence of conjunctive capability dependencies: two agents each individually inca- pable of reaching any forbidden capability can, when combined, collectively reach a forbidden goal through an emergent conjunctive dependency.

arXiv comments
Journal reference
DOI
Subjects
cs.AIArtificial Intelligence

On Hacker News

Linked by 1 identifiable commenter in 1 comment across 1 thread, from Apr 2026 to Apr 2026.

Comments linking this paper, per year.

Discussed in comments1 thread

  1. Show HN: Endo Familiar, an O-cap based JavaScript agent sandbox 1 comment

    • I wanted to ask perhaps an impossible question on this notion that I've been chewing on — particularly around the composition of "capabilities" (LLM sense of the word) and what they could acquire through some transitive closure over their "capability set(s)" in a multi-agent…

      hierophantosView on HNParent context

Submitted to Hacker News

Connected through shared HN threads, not inferred similarity or a recommendation.