‘How Long Do We Have?' Nithin Kamath On AI Threat To India's White-Collar Jobs

News18

06-06-2025

Last Updated:
New information shared by Zerodha co-founder Nithin Kamath shows a clear warning: Artificial intelligence can now handle longer and more complicated tasks than ever before.
Artificial intelligence isn't merely progressing; it's accelerating at a breakneck speed, posing a significant challenge to India's service-based economy. According to new insights from Zerodha co-founder Nithin Kamath, AI systems are increasingly capable of managing longer and more intricate tasks than ever before. This rapid evolution means many jobs traditionally performed by humans could soon become obsolete. It raises urgent questions about what the future holds for employment in the country.
On X (formerly Twitter), Nithin Kamath posted two images – one chart shows that the length of tasks AI can complete at a 50 per cent rate is doubling every seven months. Just a few years ago, earlier models like GPT-2 were limited to handling only brief, simple queries lasting seconds.
Today, GPT-4o can sustain performance for nearly an hour. The most advanced systems, such as Sonnet 3.7, are already managing complex, multi-hour assignments like training image classifiers—tasks that have moved beyond experimental phases and are becoming true substitutes for human work.
A second, more concerning chart tracks AI's accuracy across different task lengths and complexities. For short and simple tasks, leading models now achieve over 90 per cent accuracy. Even more complicated and ambiguous challenges, once thought to be beyond AI's capabilities, are now seeing significant improvements. Only the most difficult category—long, complicated tasks—still lags with less than 30 per cent success.
However, that gap is steadily closing.
One user remarked, 'It's hard to predict, but it feels like we're nearing a major turning point."
Another added, 'Not much time left, but ultimately this progress will improve our lives. It could lead to sustainable abundance and even communism—after all, capitalism is paving the way."
A different commenter emphasised the importance of adaptability, saying, 'As long as people keep reskilling and adding value to their organizations, they'll stay relevant."
Meanwhile, someone else reassured others by pointing out, 'There's still plenty of time because AI relies heavily on human knowledge as middleware."
The discussion highlighted a mix of concern and cautious optimism about the future impact of AI.
First Published: