‘How lengthy do we now have left?’: Nithin Kamath flags looming AI disaster for India’s white collar employees
AI is not simply evolving, it’s accelerating at a tempo that threatens to overrun India’s service-based economic system. New information shared by Zerodha co-founder Nithin Kamath paints a stark image: synthetic intelligence is now able to performing longer, extra complicated duties than ever earlier than, and it’s bettering so quick that many human jobs could not stand an opportunity.
One chart reveals that the size of duties AI can execute with 50% success is doubling each seven months. Only a few years in the past, fashions like GPT-2 might deal with seconds of fundamental queries. At the moment, GPT-4o can maintain its personal for practically an hour.
And essentially the most superior fashions—like Sonnet 3.7—are already dealing with multi-hour jobs reminiscent of coaching picture classifiers. These are not experiments. They’re replacements.
A second chart is extra unsettling. It maps AI’s success throughout job issue and period. On clear, quick duties, prime fashions at the moment are attaining over 90% accuracy. Even messier, ambiguous issues—as soon as thought of past AI’s attain—are seeing sharp positive aspects. Solely the toughest class, lengthy and messy duties, stays beneath 30% success.
However even that barrier is eroding.
Confronted with this information, Kamath posed a sobering query on X: “How lengthy do we now have left?”
He’s not alone in elevating the alarm. Atomberg founder Arindam Paul estimates that 40–50% of white-collar jobs in India—particularly in IT and BPO—might vanish.
A current IIM Ahmedabad research reveals 68% of white-collar professionals consider AI will automate their jobs, at the very least partially, inside 5 years. Jefferies forecasts that half of all entry-level white-collar roles—from advertising to gross sales to software program—could also be worn out in simply 1–5 years.
This isn’t theoretical. It’s taking place. And India, with its heavy reliance on companies and restricted manufacturing capability, is uniquely susceptible.
