Inside Apollo: Science of Scheming & AGI safety products

​Why does scheming get worse as AI models get smarter?

​Join Apollo Research and BlueDot Impact for a behind-the-scenes look at Apollo's new research agenda to build a science of scheming. Their recent work shows that as models become more capable, they also become better at covertly pursuing goals their developers didn't intend. Training can reduce this, but doesn't eliminate it.

​Since our first Inside Apollo event, Apollo has become a Public Benefit Corporation, raised a seed round led by 50Y and opened a new office in San Francisco.

​​Join us to learn:

• ​What Apollo's new research agenda (Science of Scheming) is trying to solve

• ​Why Apollo is building AGI Safety products and the ways research that addresses future AI risks can be productised right now

• ​Open roles across research, engineering, and operations in San Francisco

​​🎙️ Speakers

• ​​Marius Hobbhahn – CEO/Co-Founder

• ​Alexander Meinke — Head of Research

• ​Joping Chai — People & Operations