Nikhil Kamath’s Podcast: Anthropic CEO Dario Amodei has warned that an “AI tsunami” is on the way. He said artificial intelligence is reaching human-level abilities much faster than most people expect. Speaking on a podcast with Zerodha co-founder Nikhil Kamath, Amodei said society is not fully prepared for the economic and geopolitical changes advanced AI systems could bring.
Anthropic CEO Dario Amodei compared the current situation to a tsunami that is clearly visible on the horizon. According to him, people can see it coming but are still trying to convince themselves that it is not real. “It’s like a tsunami heading toward us. We can see it clearly, yet some people say it’s just a trick of the light,” he said.
Amodei added that AI’s impact will differ across industries. Jobs that require strong human interaction and emotional understanding may take longer to be affected. However, he noted that AI is already taking over tasks such as coding, mathematics, and scientific research.
The conversation also turned to a key question shaping AI’s future: the role of data. Amodei challenged the traditional belief that real-world data is always the main driver of AI progress. In fields such as mathematics and coding, he explained, the data is often not collected but generated. “When you train on math or coding environments, you’re not really getting data. It’s more synthetic. You’re creating the data,” he said. According to him, this kind of dynamic, self-generated data is becoming increasingly important in advancing AI systems.
Looking at the bigger picture, Amodei maintained a balanced view. He acknowledged both the enormous potential of AI and the risks that come with it. “My instinct is we’re about to cure a lot of diseases,” he said, highlighting the technology’s promise. At the same time, he stressed the need for responsible development and effective governance.
He also noted that while AI models are already handling a growing share of coding tasks, the broader responsibilities of software engineering will take more time to automate. Still, he believes full end-to-end automation could arrive sooner than many people expect. (With IANS Inputs)
