Goodfire is an AI interpretability research lab focused on understanding and intentionally designing advanced AI systems.
mech interp unicorn, one of the best places to do such research in the world today, lots of momentum
Nathan LabenzNathan's Companies you should join
- Nathan Labenz
Research, evaluations, and monitoring tools to detect AI scheming and secure frontier systems.
started as a non-profit focused on deception and scheming research (a number of interesting and kind of scary episodes with them), but moving toward commercializing their agent monitoring tech
Nathan Labenz
agent standard author and auditor – playing a critical role in nascent agent insurance market – just raised a second big round, notable customers like ElevenLabs
Nathan Labenz
AI tools for scientific research and evidence synthesis.
a favorite of mine from early on, AI-supported decision making for high stakes environments, good traction in Pharma - a contrarian play that is trying to figure out how to create value via structured, externalized reasoning while taking seriously and avoiding falling victim to the bitter lesson
Nathan Labenz
Higher confidence in aligned superintelligence.
non-profit with $150M in funding, led by one of the most accomplished people in AI history (Geoffrey Irving, most recently Chief Scientist at UK AISI), very AGI pilled, pushing the limits of what various branches of mathematical theory can say about AI systems
Nathan Labenz
non-profit with a broad AI safety agenda, from interpretability research to policy, also does pre-deployment testing and is likely to do more - aiming to double staff in the next year
Nathan Labenz
Applied AI engineering and alignment research, funded by an independent consulting business.
one of my favorite less-well-known companies, entirely bootstrapped to ~150 people, afaik, with lots of interesting stuff going on in addition to their core tech business
Nathan Labenz
