Building & Scaling the AI Safety Research Community, with Ryan Kidd of MATS

2026-01-04 · Source: The Cognitive Revolution · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Emerging Technologies & Innovation, Cybersecurity & Data Privacy · Depth: Advanced, extended

Summary

Ryan Kidd, co-executive director of MATS, discusses the evolving landscape of AI safety, AGI timelines, and the critical need for a robust research community. He highlights MATS' role as the world's largest AI safety research talent pipeline, having supported 446 fellows with 80% securing permanent jobs in the field. Kidd notes the high level of disagreement among experts regarding AGI timelines, with Metaculus predicting strong AGI around mid-2033, but emphasizes preparing for earlier scenarios. The discussion covers the blurred lines between safety and capabilities work, the increasing sophistication of AI deception, and the importance of developing technical governance solutions. MATS offers diverse research tracks, including empirical research, policy and strategy, theory, technical governance, and compute infrastructure, with applications for its Summer 2026 Fellowship closing on January 18th.

Key takeaway

For AI Scientists and Research Scientists aiming to enter or advance in the AI safety field, prioritize developing demonstrable technical skills and producing tangible research outputs. The field is highly competitive, but organizations like MATS provide critical pathways and mentorship. Focus on understanding core AI safety concepts and consider building a portfolio of projects that showcase your ability to work with current AI systems, as this will be essential for standing out in a talent-constrained market. The deadline for the MATS Summer 2026 Fellowship is January 18th.

Key insights

AI safety requires a portfolio approach, balancing technical research with governance and preparing for uncertain AGI timelines.

Principles

• All safety work is fundamentally capabilities work.
• Lowering the "alignment tax" via technical research is crucial.
• Portfolio approach to research bets is the only defensible position.

Method

MATS identifies "Connectors" (new paradigm creators), "Iterators" (empirical researchers), and "Amplifiers" (scaling research teams) to address varied talent needs in AI safety, emphasizing proficiency with AI assistance tools.

In practice

• Focus on technical skills and understanding frontier technology.
• Build tangible research output for applications.
• Consider independent projects or grant funding.

Topics

• AI Safety Research
• AGI Timelines
• MATS Program
• AI Talent Pipeline
• AI Governance

Best for: AI Scientist, Research Scientist, AI Researcher, AI Student, Policy Maker

Related on AIssential

• Open Thread 436 — Diverse efforts are underway in AI governance, research, and political advocacy, often involving major tech players.
• The inaugural Redwood Research podcast — AI safety requires pragmatic strategies, acknowledging both technical challenges and organizational realities, to manage escalating risks.
• Best of Lenny’s Newsletter 2026 — AI progress is accelerating, necessitating a proactive focus on safety and alignment to navigate societal transformation.
• Announcing the OpenAI Safety Fellowship — The OpenAI Safety Fellowship fosters external research on AI safety and alignment through mentorship and resources.
• You Are Being Told Contradictory Things About AI — Contradictory narratives dominate AI's future, spanning job displacement, AGI paths, and the necessity of recursive self-improvement.
• Washington weighs new bipartisan AI proposal — Rapid AI advancement necessitates urgent, multi-faceted regulatory and safety measures, alongside addressing infrastructure challenges.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by The Cognitive Revolution.