“It will not be easy,” says Curefit founder

Photo of author

By [email protected]


Mukesh Bansal, the founder of Curefit and Myntra, questioned the gap between the rapid progress of the large language models (LLMS) and its impact in the real real world below.

In a post on X (previously on Twitter), Bansal wrote, “Everyone builds AI agents, however, a few artificial intelligence agents in production are doing real work. LLMS continued to improve and better, however, P & LS does not change through basis points! What happens?”

He stressed that the construction of “production category agents” requires the basic rethinking of commercial operations, and not only to automate the current workflow with claims. “The agents company is not similar to the traditional software company or platforms at all. It is an intersection between MCKINSEY and Infosys … it may be good for upgrade but it is bad to work.”

Bansal noted that artificial intelligence agents at the level of institutions are demanding the best human talent that works alongside LLMS to create solutions that can “develop dynamically” and reshape industries. He added: “If this is easy, the agents will already operate the world! The revolutions are supposed to be easy, even if the inevitability is like this.”

In a separate publication, Bansal warned that the ecosystem of deep science and technology in India needs urgent strengthening in light of global geopolitical tensions and upcoming artificial intelligence wave. “This is not an invitation to wake up. This is existential,” he wrote, calling for further investment in research that depends on curiosity, global infrastructure, and efforts to attract senior Indian scholars to the homeland.

“The twenty -first century requires nothing less. Vixet is not possible without this,” he stressed.





https://akm-img-a-in.tosshub.com/businesstoday/images/story/202508/6898a4dd8d2a7-he-stressed-that-building-production-class-agents-requires-a-fundamental-rethinking-of-business-pr-10553664-16×9.jpg

Source link

Leave a Comment