The past 2 years has seen breathless proclamations by the media and ‘experts’ of an impending or inevitable ‘white-collar job apocalypse’ due to AI:
With the launch of DeepSeek, we can expect such headlines will continue. My take is, the eagerly-awaited white-collar job die-off is motivated more by wishful thinking than economic reality. Claims by the media of the imminent untold destruction of millions of white-collar jobs are unfounded. To the contrary, I predict the professional and creative-class will continue to thrive.
Even though AI can produce code, someone still has to be taught to use the AI, which means a job. Or to turn the AI code into something that is a useable product, or to combine the code into the finalized result. AI is thus more analogous to assembly or machine code underneath the hood. Too many people talk about AI like it’s all or nothing. It’s more like AI makes existing workers more productive. It does not mean masses of people are rendered destitute.
Moreover, some of these claims betray an understanding of basic economics. Jobs are always being replaced and created. To say “millions of jobs will be destroyed by AI,” but ignore that other jobs will be created or that job creation and destruction is part of the normal functioning of an economy, is irresponsible reporting by the media.
I predict there will still be top-dollar paid to top talent in spite of AI. Those with sought credentials from top schools, indicative of high IQ, will continue to command a premium in the labor market. Being smarter makes it easier to adapt, even to AI. This is what the doomsayers get wrong: It’s those who have the most talent who are the most capable of adapting to changing economic times. Even with advanced AI, the >140 IQ guy who wins math competitions and has a STEM degree is still going to be a more valuable employee. In this case, the value is in using AI optimally.
For example, engineer salaries at META and other top tech companies are the highest ever, with no sign of slowing. Despite the fear or promise of AI replacing coders, top coders still earn a ton, because AI cannot replace it all. There are still those edge cases or combining AI-generated code into a working product. Coders, for example, can be repurposed to oversee AI-generated code instead of creating the code themselves. This is what already happens, such as the use of codebases and libraries. Coders do not create everything from scratch when starting a new project.
It’s hard to think of any white-collar profession or industry that has seen a marked decline in headcount or compensation directly attributable to AI. In fairness, two examples that come to mind are Chegg and Getty Images, but even at their peak these companies were tiny relative to today’s tech and finance behemoths. Even Adobe, the maker of Photoshop, has managed to to thrive even though there are cheaper AI alternatives to its pricey proprietary photo editing software.
Same for top authors on Substack, who as of 2025 have seen record earnings and readership despite AI. If there is any profession that is vulnerable to AI, writing should be high on that list, yet as far as Twitter and Substack are concerned, things are booming. AI cannot readily reproduce readership or branding, the subtle nuance of the author’s ‘voice’, or the type of content that readers find engaging, except perhaps a crude facsimile. Of course, AI can produce competent prose, but success at writing is much more than that.
On the other extreme, any job that involves interacting with people will also thrive. This is includes mundane food jobs. The expectation during Covid was that many in-person jobs would be rendered obsolete, but these jobs have only exploded since then. Fast food, retail, hospices, financial advisors, banking, salons, childcare, dentistry, colonoscopies, OBGYN, etc.–these are hard or impossible to automate.
Moreover, as the world becomes wealthier due to AI and improved productivity, demand for such services will only increase. Door Dash for example is a huge deal now, with growth surging since the pandemic. High-income people love using Door Dash for its convenience and privacy despite it being much more expensive. This what happened during and after Covid, in which the US economy flooded with stimulus money underwent a boom, and then this led to huge demand for service sector and blue collar jobs, such as home remodeling.
Even with the advent of robo-advisors, retiring boomers and gen-x will see huge demand dealing with more complicated matters like estates, inheritances, and annuities that will require in-person meetings. After my paternal grandma died, it was was a huge headache involving many accounts and having to decide how to handle it. My grandad died 13 years earlier and left it to her, but she didn’t have any clue or inclination of how to manage it.
I think some of this doom and gloom is motivated by a desire for AI to create a more equitable society, to reset it to a more equitable state, or to disrupt old industries. But industries adapt rather than are overthrown. Every trillion-dollar tech company is now also effectivity an AI company, having invested heavily in AI infrastructure. Rather than a monopoly or duopoly, like Google or Facebook, AI has many more players or niches instead of being so lopsided or winner-take-all. Look how DeepSeek came out of seemingly nowhere and threw the AI and GPU status quo into disarray overnight. This has never happened with Google, Microsoft and the rest of old tech guard.
This is not to say there won’t be disruptions, but such bold claims can be dismissed for a want of evidence. Even if some white collar jobs are replaced by AI, skilled workers are still in a better position to adapt compared to unskilled workers, due to having better credentials. It’s time to give the sensationalism a rest.