"Figure Out How To Stay Relevant": Sridhar Vembu's Warning To Software Engineers As AI Takes Over Coding

Artificial intelligence could be on the verge of solving a huge share of advanced mathematical problems, and the shift may happen much sooner than many expect. In a post on X, Zoho founder Sridhar Vembu believes AI could potentially make major breakthroughs in mathematics within months. Vembu said people may need to accept the possibility that AI could eventually tackle almost all of mathematics, even if the idea is unsettling. 

"We have to accept the possibility that all mathematics could fall to AI within some months, however unsettling that might seem," he wrote 

He pointed to the rapid progress AI has already made in areas that resemble games such as chess and Go.

A bigger challenge, he argued, is coming up with entirely new mathematical ideas or frameworks that humans would find meaningful and interesting. "The "humans would consider interesting" could be considered a fundamental limit of AI or merely a face-saving formula to protect human ego, but humans set the rules (for now!)," he added. 

Vembu also linked this development to programming. He suggested that AI could eventually produce a large portion of software code both efficiently and correctly. Advances in formal verification systems such as Lean, which are already used to check mathematical proofs, could help address concerns about whether AI-generated code is reliable. Such systems could increasingly be used to verify software written by AI.

See the post here:

"We have to accept the possibility that all mathematics could fall to AI within some months, however unsettling that might seem."

The part of mathematics that is like playing Chess or Go, coming up with ever cleverer moves, will be conquered by AI. That is now evident.

The part… https://t.co/PWsfS4PtOL

— Sridhar Vembu (@svembu) September 6, 2026

However, Vembu stressed that the future is unlikely to be about AI completely replacing humans. People will still be involved in the process, but the more important question may be how many humans will actually be required.

"The point is not that "AI will do it all". There will be humans involved in the process. The key question is "how many humans are needed?", and the answer right now is sobering and unsettling. All of us in software need to figure out how to stay relevant. Let me get back to work. Please wish us luck," he concluded the tweet.