Dario Amodei's conversation on Nikhil Kamath's podcast WTF Is is turning heads in the artificial intelligence world. The interview between the Anthropic CEO and Zerodha co-founder ran for a little over an hour and split open Anthropic's philosophy to navigate an upcoming tsunami of human-level AI.
Amodei questions how a society, which is so close to reaching human-level intelligence, can be so unprepared for it? According to the Anthropic CEO, the society is currently in denial about the speed of AI progression. He notes that while the technical side (scaling laws) is working perfectly, the societal awareness is lagging dangerously behind.
Amodei applied the 'trick of light' analogy to explain public's lack of awareness regarding AI risks.
"It's as if this tsunami is coming at us. It's so close, we can see it on the horizon, and yet people are coming up with explanations like, 'Oh, it's not actually a tsunami, it's just a trick of the light.' I think there hasn't been a public awareness of the risk," Dario remarked.
At a juncture where almost every profession feels endangered because of the rise of AI, Amodei broke down that coding will be the first to get automated.
The Anthropic co-founder made a bold prediction. “I think coding is going away first,” he said, adding that while AI systems are rapidly improving at writing software, engineering divisions such as architecture, product sense and user understanding will take longer to automate.
Artificial intelligence is set to automate coding before it replaces broader software engineering work, Amodei said, outlining how AI will reshape skills and jobs.
At the same time, Dario emphasised on the importance of human inputs and existence even in a pro-AI world. Pointing to the challenge of an expanding automation scope, the value of other moats like physical-world integration, human relationships, and institutional knowledge may surface more.
"As AI automates technical tasks, human-centric tasks may grow in importance," he said. He argued that comparative advantage still matters and said that even if humans do 5% of a workflow, AI amplifies that.
Amodei warns entrepreneurs against building "thin wrappers"— products that simply repackage an AI model without adding unique value. In a world where frontier models like Claude are rapidly evolving, a business built on a simple interface can be snapped out of existence overnight by a single update. To survive, companies must build deep "moats" rooted in domain expertise, complex industry integration, and unique workflows.
From a long-term perspective, Amodei believes Biotech will likely see a renaissance driven by AI. Key areas of focus include mRNA-like programmable biology, peptide therapies, and cell-based therapies (e.g., CAR-T).
Amodei's embers for exploration in AI sparked from a despair around how complex biology is. "The RNA gets spliced in different ways depending on where it is in the cell. Then it gets post-translationally modified, phosphorylated, combined with other proteins. I started to despair that it was too complicated for humans to understand," the former biologist recalled.
Later, he noticed early work around one of the first neural nets called AlexNet almost 15 years ago and remembered thinking AI is starting to work.
"It has some things in common with the human brain, but has the potential to scale much larger and learn tasks like biology.
Maybe this is ultimately the solution to solving biology," he said.
In the end, Amodei advised the younger generation to focus on human-centred tasks, physical-world interfaces, analytical skills, and "most importantly", on critical thinking.
Essential Business Intelligence, Continuous LIVE TV, Sharp Market Insights, Practical Personal Finance Advice and Latest Stories — On NDTV Profit.