Anthropic CEO Dario Amodei Reveals The Reason Behind His OpenAI Exit

Anthropic CEO and co-founder Dario Amodei has shed new light on one of the biggest questions in the AI industry: why did he leave OpenAI and start a rival company? Speaking on investor Nikhil Kamath's podcast WTF Is, Amodei revealed that two core beliefs ultimately pushed him to walk away from OpenAI in 2021. One centred on the future potential of AI itself. The other was about how safely that future should be built. His comments offer a rare glimpse into the thinking that helped shape Anthropic, the company behind Claude, which today competes directly with ChatGPT.The Idea That Changed EverythingAmodei joined OpenAI in its early years and eventually became the head of research. According to him, the turning point arrived around 2019 during

the development of GPT-2. At the time, he became convinced that simply increasing the size of AI models, feeding them more data and giving them more computing power could unlock dramatic jumps in capability."You find incredible increases in performance," Amodei said while describing what happens when models are scaled up.Today, that idea sounds obvious. Back then, it wasn't. According to Amodei, many people both inside and outside OpenAI were sceptical about how far scaling could take AI systems. He and a small group of colleagues believed otherwise and pushed that view internally as models like GPT-2 and GPT-3 evolved.Anthropic Co-Founder Warns AI Needs A Brake Pedal As Race AcceleratesA Growing Concern About SafetyBut model performance wasn't the only thing on his mind. As AI systems became more capable, Amodei began thinking more seriously about their long-term impact. He believed advanced AI could eventually become a general-purpose cognitive tool capable of transforming industries, economies, and even geopolitics."The economic implications are gonna be enormous. The geopolitical implications are gonna be enormous," he said. "It's gonna transform how the world works." That belief led to a second conclusion: AI safety could not be treated as a secondary concern.Why He Decided To LeaveAmodei explained that while OpenAI talked extensively about responsible AI development, he wasn't fully convinced the organisation was approaching safety with the level of commitment he felt was necessary. Rather than continue debating the issue internally, he chose a different path. "Don't argue with someone else's vision," Amodei said. "If you have a strong vision and you share that vision with a few other people, you should just go off and do your own thing." That's exactly what happened.The Birth Of AnthropicIn 2021, Amodei left OpenAI and co-founded Anthropic alongside his sister Daniela Amodei and several former OpenAI researchers. Since then, the AI startup has grown into one of the industry's biggest players. Backed by billions of dollars in investment from Amazon and Google, Anthropic's Claude models now compete head-to-head with ChatGPT across both enterprise and consumer markets.