Koh Boon Hwee: AI Means Engineers Must Think Better, Not Code Better

Singapore tech and investment veteran Koh Boon Hwee has offered a characteristically frank assessment of the AI landscape, declaring that artificial intelligence will fundamentally transform what it means to be a software engineer — and that most AI infrastructure investments remain loss-making for investors. In an exclusive interview with The Business Times, the early Razer backer argued that the benchmark for engineers is shifting from 'I can code better than you' to 'I can think better than you.'

Koh, who visits Silicon Valley every other month, described the US tech corridor as having 'a unique combination of money, entrepreneurs, and most importantly, a risk-taking attitude.' But he is steering clear of AI infrastructure plays, calling them 'loss-making for investors' given the billions required to build and train large language models. Instead, his focus is on companies that use AI infrastructure to improve operations, particularly in specialised agentic AI systems that gather data, design objectives, and execute plans autonomously.

On the impact for Singapore's tech workforce, Koh predicted that coding will shrink to under five percent of a software engineer's time as AI engines handle routine code generation. 'If I were a software engineer, I'd go back to university and improve my user needs-testing skills and worry less about coding,' he said. He described traditional examinations as 'the easiest way — and also the most erroneous way — of evaluating future skills,' arguing that an education model that worked for the last 50 years will not work for the next decade.

The sectors seeing the earliest AI dividends, according to Koh, include telcos and airlines with massive data-processing needs, healthcare where AI cross-checks drug interactions, insurance for processing thousands of validity claims, and banking for compliance checks. On chatbots, he noted that US customer satisfaction has 'gone up significantly' with advanced AI agents that recognise roadblocks and hand off to humans, learning after 10 to 20 interactions to take over again. But he cautioned that AI still cannot judge customer importance, assess vendor trustworthiness, read body language, or handle edge cases with insufficient data — skills that remain uniquely human.

Why it matters for Singapore: Koh Boon Hwee's views carry weight in Singapore's tech investment community — he was an early backer of Razer and chairs Altara Ventures. His warning that AI infrastructure is overhyped for investors while the real value lies in application-layer innovation is a counterpoint to the current frenzy around GPU clusters and data centre builds. For Singapore's engineering talent, his call to shift focus from coding proficiency to higher-order thinking, empathy, and communication skills signals a needed reorientation of how the city-state prepares its tech workforce for the AI era.