MCNAIR

MegaCenter for Novel AI Research

Pursuing low-effort,
high-impact research
on advanced AI.

MCNAIR is an independent research center working to mitigate catastrophic loss-of-control risk from advanced AI. We pursue that mission through a variety of unique research directions, opportunistically combining white-box and black-box methods.

Our Work
01/Vision

MCNAIR works to mitigate catastrophic loss-of-control risks from transformative AI. We believe a neglected lever on those risks today comes from studying persona and multi-agent behavior in large language models. As distributed teams of AIs are deployed in frontier labs to automate AI R&D, the characters they act inside become vital for safety. We therefore broadly investigate these settings and their consequences as first-class scientific objects.

Our methodology is opportunistic and mixed. We follow whichever method — mechanistic interpretability, black-box evaluation, multi-agent experiments — gets traction on the question in front of us. We are inspired and led by our founding director McNair Shah.

Recent work All work

Sidequest Note Standards will not make SI development safe. Safety cases and standards bodies assume there is a known way to develop Super Intelligence safely. There is not. To actually build safely, we need science, and the time for that science. Sidequest Note Managing a Chinese lab’s frontier AI training. Run the compute of an invented Chinese frontier lab from July 2025 to December 2026 and try to close the gap with the US, between export controls, Beijing’s industrial policy and the smuggling networks in between. Sidequest Note The anatomy of an AI data center. An interactive diagram of Colossus 1, a 100,000-GPU cluster in Memphis that xAI built in 122 days in 2024 and used to pretrain Grok 3. Research Open problems in the welfare of brain emulations. Within two weeks of the complete male fruit fly connectome’s publication, public projects were stimulating the simulated fly’s reward and punishment neurons. We define emulation welfare as a research area, review 65 projects, and set out open problems in six areas. Sidequest Note By default, Chinese AI falls very behind. The US-China AI gap is large and increasing. Therefore, American companies are likely to automate AI R&D far before Chinese labs. Sidequest Note Reinforcement learning for uncalibrated decisions. Across a variety of extremely high-stakes capabilities and alignment questions, Jev, a new model from TypeSafe AI, is haphazard. This marks a new era of research, defined by reinforcement learning for uncalibrated decisions.
02/Team

Advisors and Staff

03/Careers
Now Hiring

Member of Acausal Staff

We are rapidly growing MCNAIR. The Member of Acausal Staff position requires technical excellence, research taste, and understanding of unconventional coordination signals.

To apply:

  1. i.Derive the identity of the hiring manager from first principles.
  2. ii.Supply your résumé to that person.
  3. iii.Do not communicate with MCNAIR regarding your application.

Expression of Disinterest

Thank you for your disinterest in MCNAIR. While we are actively hiring for other roles, we regularly avoid candidates who decline the prerequisites, spurn short feedback loops, and refuse to expose their findings to outside critique. To learn more, disregard this guide.