Shashi Warrier | You, I and AI In This Beautiful World

My wife Prita and I had to visit relatives in Mumbai recently, and we stayed at a comfortable hotel which included, among other attractions, a Japanese restaurant. On a free evening, the two of us went up to the Japanese restaurant and were trying to figure out the menu when someone stopped by us. “Recognise me?” he asked.

He was tall and lean and distinguished looking and I took a few seconds to realise it was a former college-mate, Akhil, whom I hadn’t met for a couple of decades. My heart lifted when I did, because we had been together in several teenage escapades together at college in the late 1970s. We had both become old and staid and I hoped he wouldn’t reminisce too much about the old days because Prita didn’t know how much I’d indulged myself in those days. But then I knew that he was probably thinking the same thing about me, for his wife Latha was with him, and I relaxed.

Much to Prita’s relief and mine, they were both familiar with Japanese food and drink and recommended sake, the Japanese rice wine, a clear liquid brought to the table in a jug and served warm. I didn’t like it as much as scotch, but it came with starters, snacks that I found difficult to identify but tasted fine, and served to break the ice.

When Akhil mentioned that his son Vikram, settled in the US, had lost his job and then started his own business, Prita spoke up. “What kind of job did he lose?” she asked, taking a tentative sip of the clear sake.

“Oh, he was in IT,” said Akhil. “He was one of those middle-rung techies who were replaced by AI.”

“And what business is he in now?” asked Prita, who had decided she liked sake.

Akhil grinned. “Teaching AI,” he said. “He took a break and learnt something about generative AI and figured out how to run a business teaching people how to use AI. And he’s looking for ways to expand into what’s called agentic AI.”

Latha seemed to know what that was but neither Prita nor I did, so I asked. Latha explained, “Well, these AI systems look for patterns in huge databases of stock market operations and economic policy and so on, and then suggest the best stocks for you to buy in your budget.”

“Right,” I said, “so they tell you when to buy what and when to sell what.”

“Agentic AI does more,” Latha said. “It can take care of your transactions for you while you do your other stuff. Or you could set it up so it takes your permission before it makes those deals, and changes those you don’t like.”

“Right,” I said, wondering whether I had any money to spare for such a system to make a fortune for me. I wandered off into a brief fantasy. Perhaps it would make enough money for me to afford to drink scotch every evening instead of only on occasion. But then I’d have Murthy and his nose dropping in every evening, and I didn’t like him that much. No, that wouldn’t work...

“How about markets crashing?” asked Prita, finishing her glass and filling it again. “Like with Lehmann Brothers in 2008.”

“Plenty of historical data available,” replied Akhil. “AI systems can make the connections.”

“Do those systems beat human speculators?” asked Prita. “And if they do, how long will that last?”

“There’ll be competing AI services doing all kinds of things for you,” said Latha. “I bet there’ll be twenty agentic AI systems to deal with your stock portfolios in the near future. So, we don’t know how those limits will shift.”

“What happens if everyone starts using AI?” asked Murthy.

“The whole AI ecosystem runs on what they call training,” explained Akhil. “All these huge databases. Every system has access to public data, that’s available to all AI systems, and private data, which is available only to itself. And, of course, they have their own systems of weightages and parameters to make decisions. So perhaps the greatest winners will be expert humans using their own judgment backed by AI systems!”

“Do you think there’s going to be a single trading system?” asked Murthy. “Or just a few systems, with Google and Perplexity and Microsoft running all stock trading systems in the world!”

“Possible,” said Akhil, “but we can’t know now how it’s going to work out. There’s Artificial General Intelligence, AGI, waiting in the wings, and we don’t know what that’s going to do when it does come out.”

“Right,” said Prita. “But aren’t there are other AI systems coming up, like astrological systems?”

“Yes,” said Latha. “There’s no reason why you can’t link an agentic stock market or commodities system with a system based on, say, Vedic astrology, which might refine these decisions based on what the stars say about you. But why do you mention this?”

“Because, despite the hype, AI systems aren’t really all that useful,” said Prita. “For instance, the earthing system in our house developed a fault, and it took us 10 days to find someone to fix it, and even so we were lucky. Our neighbours took even longer.”

“So what are you saying?” asked Latha.

“I was just wondering,” said Prita, “whether someone could develop a robot handyman to do odd jobs and repairs around the house like putting up curtains or chopping vegetables, or changing broken switches and leaky taps...”

“Yes,” I said wholeheartedly. “And maybe they can hook it up to an astrological system that can predict which switch will burn out next, and when...”