Truly AI Native Companies with Sebastian Graf

Episode notes

Sebastian Graf explains the coming rise of AI native companies. He shares four safety pillars for autonomous firms: code freeze, CICD compliance, regulatory model training, and harm-reduction finance. The episode mixes technical detail, a real Anthropic experiment, and broad governance concerns about trust and social impact. Top 3 Key Learnings AI native company pillars: Four safety pillars: code freeze, CICD, regulatory RL training, and harm reduction mechanisms. Test and freeze code: Freeze codebases and run automated tests to ensure predictable behavior and regulatory compliance. Trust via cost and trials: Lower cost and stepwise use build trust; people try low-risk services first then adopt more. About Sebastian Graf Sebastian Graf describes himself as an engineer by profession, but an educator at heart. In his work as an engineer, he is driven by the belief that technology should lift everyone, enabling people to live extraordinary lives in extraordinary ways. He acknowledges that this is challenging in a world where many technologies are built with incentives that divide society, exploit the environment, and widen wealth inequality.