Job Paradox
I've had experience working with companies that are all-in on AI and want every last person to skill up. There is ample support and opportunity to reinvent one's job, be recognized and rewarded for it. Sadly, their offshore teams in India were the most resistant to change and continued to follow their old ways of working unless forced to comply. There is a pervasive fear of job loss due to AI so folks refuse to use it and consequently do not develop the skills they need to survive it. So reading this story about how India will magically adapt and win with AI felt out of sync with reality on the ground.
The "AI job paradox" sounds like a nice solution to a messy problem. The theory is simple: as machines take over routine tasks, they free up humans to focus on high-level work—empathy, judgment, and complex strategy. It’s a comforting vision for India’s workforce, but it ignores several hard realities.
Automation doesn't just change tasks; it deletes them at scale. Much of the process drag in organizations are a result of needing many people cross functions to work in co-ordination to get something one. This is mostly a relic of the past. One person managing an AI tool can often produce what twenty workers used to. Even if "higher-value" roles appear, they won't appear in the millions. For a country that needs to create a million jobs every month just to keep pace with its population, a net-loss in total roles is a crisis, not a "shift."
Then there is the reskilling myth. The idea that a data-entry clerk can simply pivot into an "ethical judgment officer" is optimistic at best and delusional at worst. These high-value skills—like nuanced decision-making—are usually the result of elite education and many years of experience. We aren't looking at a workforce moving up; we are looking at a workforce being hollowed out. The top 5% will thrive as AI orchestrators, while the rest are pushed into low-wage manual labor that remains too cheap to automate.
The article claims humans are indispensable because AI makes mistakes. A lot of people cling to this idea very passionately and refuse to entertain any other possibilities. But history shows that companies often accept "good enough" results if they save money. We’ve already seen this in customer service and content moderation. Instead of being empowered "orchestrators," many workers will likely become "rubber-stampers," clicking "approve" on machine-generated work for lower pay. I've had tenured engineers tell me that catastrophe will follow if we use AI to build and mantain API inegerations for example. They have chosen not to understand why infact this is one for the ideal use cases for AI; the so called spec-driven autonomous loop. It would be nice to see it attempted instead of declaring code red at the very thought of it.
The "AI job paradox" sounds like a nice solution to a messy problem. The theory is simple: as machines take over routine tasks, they free up humans to focus on high-level work—empathy, judgment, and complex strategy. It’s a comforting vision for India’s workforce, but it ignores several hard realities.
Automation doesn't just change tasks; it deletes them at scale. Much of the process drag in organizations are a result of needing many people cross functions to work in co-ordination to get something one. This is mostly a relic of the past. One person managing an AI tool can often produce what twenty workers used to. Even if "higher-value" roles appear, they won't appear in the millions. For a country that needs to create a million jobs every month just to keep pace with its population, a net-loss in total roles is a crisis, not a "shift."
Then there is the reskilling myth. The idea that a data-entry clerk can simply pivot into an "ethical judgment officer" is optimistic at best and delusional at worst. These high-value skills—like nuanced decision-making—are usually the result of elite education and many years of experience. We aren't looking at a workforce moving up; we are looking at a workforce being hollowed out. The top 5% will thrive as AI orchestrators, while the rest are pushed into low-wage manual labor that remains too cheap to automate.
The article claims humans are indispensable because AI makes mistakes. A lot of people cling to this idea very passionately and refuse to entertain any other possibilities. But history shows that companies often accept "good enough" results if they save money. We’ve already seen this in customer service and content moderation. Instead of being empowered "orchestrators," many workers will likely become "rubber-stampers," clicking "approve" on machine-generated work for lower pay. I've had tenured engineers tell me that catastrophe will follow if we use AI to build and mantain API inegerations for example. They have chosen not to understand why infact this is one for the ideal use cases for AI; the so called spec-driven autonomous loop. It would be nice to see it attempted instead of declaring code red at the very thought of it.
I was chatting with an offshore engineering manager recently ans he shared how difficult its been for him to bring his team along because they code manually and do not want to let go of their ways of working because it effaces their professional identity. He is trying to motivate them by saying they have the choice of falling in line or being replaced with one senior dev in America who runs a team of agents the same size of the current team in India. The compensation math would work out. His team has understably pushed back saying what is to prevent that from happening already. These are real issues and there are no good answers. He has been able to bring half of his team along to the new ways of working so far.
AI moves at the speed of light. Indian education and corporate training move at the speed of a glacier. By the time our training programs catch up to the "skills of the future," the technology will have moved again. The paradox is that the more "human" a job becomes, the more exclusive it becomes. Unless we address the massive disparity in training and the raw loss of volume, the AI wave won't lift all boats—it will just widen the gap between the shore and the ship.
AI moves at the speed of light. Indian education and corporate training move at the speed of a glacier. By the time our training programs catch up to the "skills of the future," the technology will have moved again. The paradox is that the more "human" a job becomes, the more exclusive it becomes. Unless we address the massive disparity in training and the raw loss of volume, the AI wave won't lift all boats—it will just widen the gap between the shore and the ship.