AI coding has hit an 'intelligence saturation' point, Cognition cofounder says

For years, the AI competition has preferred the smarter model.

Russell Kaplan, the chairman and cofounder of AI coding startup Cognition, the business rearward software-engineering phase Devin, stated that could be changing for a expanding figure of software-engineering tasks.

"Many tasks inside application engineering are now 'intelligence saturated,'" Kaplan wrote in a article on Monday. "Once any example can do the project well, disbursal and speed are all that matter."

Kaplan was responding to an part from The Information. Their reporting established that a 10-point advantage in example intellect now delivers fair 20% additional developer traffic, downward from nearly 69% in delayed 2025.

He stated that narrative is "directionally consistent" alongside what is happening inner Cognition.

That could be a sizable strategic change for the AI companies fighting to win application developers and endeavor contracts. OpenAI, Anthropic, Meta, Google, and a growing crop of AI labs have all highlighted their increasingly capable coding tools as a logic to choose their models.

Now, as those AI agents get improved at regular tasks, one model's slightly smarter method could matter small to users if another example is cheaper and faster.

Kaplan didn't immediately react to a petition for comment from Business Insider.