Kamal Ghorui
Code and coding are no longer moving in tandem—and this gap lies at the heart of AI’s disruption of the software services industry. “Today, we are seeing something similar—what you could call cognitive arbitrage—where labour is increasingly being replaced by machines,” former Cognizant CEO Francisco D’Souza told TOI in a recent interaction.D’Souza founded tech investment firm Recognize a few years ago and raised $1.7 billion for its second fund, just four years after closing its $1.3 billion debut. The firm focuses on companies valued between $50 million and $500 million, leveraging a partnership-driven model to accelerate growth.In his white paper, he describes this shift as the “great decoupling,” where the value of code is rising sharply even as the act of coding becomes increasingly commoditised. This creates a striking paradox. On one hand, the value of code has never been higher—evident in the extraordinary valuations of companies like OpenAI and Anthropic, driven largely by software. On the other, the value of coding—the act of writing code—is trending towards zero, as machines generate code at near-zero marginal cost. “The task of coding is becoming cheap, even while the output, well-functioning code, is worth more than ever. I call this the paradox of value,” he wrote.
However, abundant code does not equal reliable software. “More code increases complexity, risk, and accountability challenges,” he said, arguing that firms must rethink commercial models. “Input pricing is time-and-materials. Output pricing could be charged per member per month. The industry has tried outcome pricing before, but it’s hard to measure and attribute. Output pricing is a good middle ground—it lets you capture productivity gains while aligning with customer value.”Capturing this shift will require firms to move beyond legacy metrics and rethink how they create and deliver value. It also demands a different workforce. Upstream roles include value orchestrators, who align builds with business strategy; enterprise architects, who ensure integration with complex systems; and intent curators, who translate business needs into precise AI-executable specifications.Downstream roles like, results orchestrators ensure AI-generated components come together into reliable outcomes; ethics stewards embed safety and guardrails; and accountability stewards own the output, bridging the “certainty gap” when things go wrong.This transition, D’Souza said, will come with a J-curve effect. “In the short term, productivity gains from AI will reduce the need for certain types of work, creating deflationary pressure. But over time, new types of work will emerge—driven by increased technology adoption and new market opportunities.”
