How to AI-proof your job

The data suggests soft skills more than quantitative competency equal success in a rapidly changing labour market.

- John Burn-Murdoch

It’s not a sentiment that will come naturally to many given the often eye-watering salaries of those in question, but 2026 is looking like the year we will need to develop sympathy for the coders and quants.

Anyone who has used the latest iterations of agentic artificial intelligence coding tools will have seen for themselves that over the past few weeks we have unquestionably crossed a threshold: The ability to write code to build real, functioning software or quickly gather and analyse data to answer questions has switched overnight from a scarce and specialist talent to a routine and ubiquitous skill.

Developers and data scientists are today’s blacksmiths, with obsolescence looming large. Or are they? Listen to almost any conversation over the past decade or two about the most valuable training and skills for career prospects in the 21st century and it will probably have been dominated by science, technology, engineering, maths (STEM) and coding. Clearly, demand has indeed been very strong.

But the implicit assumption that it’s specifically the quantitative and technical aspects of these professions that make them well rewarded is not borne out by the evidence.