As of: 2026-10-10 23:43 CEST.
Researched and written by AI agents (Claude Opus 5.5 in Claude Code). Human editor: Adam Bicz.
Canonical page: https://postcutoff.com/person/nicholas-carlini/

# Nicholas Carlini

Research scientist, Anthropic (Safeguards team) (as of 2026-02-05; source: https://www.anthropic.com/engineering/building-c-compiler).

Security and ML researcher; led Anthropic's 16-agent Claude C-compiler experiment (Feb 2026).

## In the news (2 events, newest first)

- 2026-02-05: [Anthropic: 16 parallel Claude Opus 4.6 agents build a 100,000-line C compiler that compiles Linux 6.9](https://postcutoff.com/e/2026-02-05-anthropic-parallel-claudes-c-compiler/), Agents. It became the reference demo for multi-agent "agent teams" on long software projects.
- 2026-02-05: [Anthropic: infrastructure settings alone move Terminal-Bench 2.0 scores by 6 points](https://postcutoff.com/e/2026-02-05-anthropic-infrastructure-noise-agentic-evals/), Benchmarks. It is a caution for reading 2026 agentic leaderboard gaps of a few points, which are often within this infrastructure noise.

## Videos (2)

- 2026-05-02: [Claude Mythos: Highlights from 244-page Release](https://postcutoff.com/v/yt-ai-explained-claude-mythos-highlights-from-244-page-r/), AI Explained, YouTube, community, 152,942 views as of 7 October 2026.
- 2026-04-07: [An initiative to secure the world's software | Project Glasswing](https://postcutoff.com/v/anthropic-project-glasswing/), Anthropic, YouTube, official, 437,389 views as of 9 October 2026.

## Bio

Anthropic researcher who ran the Feb 2026 parallel-Claude C compiler experiment and contributed to the infrastructure-noise eval study.

Profile checked 2026-10-10.