A Tale of Two Mothers

Happy 250th, America

This week, I have a few unconnected thoughts.

There’s a lot being said about AI and LLMs, and how powerful they are becoming. There is a growing unease about all of this. I think it’s because a majority of people (like 90 percent of the people out there) are not tech savvy and have no way of knowing what is going on, even if it’s explained to them really well. It’s not that they’re deficient in gray matter. It’s just how tech works. Tech tends to be cult-ish, has its own jargon and lingo, and there are insiders and outsiders in this cult. In the late 1990s and early 2000s, even tech folks in software could not understand the lingo of telecom and networks, which were dramatically changing the world at that time. It’s perfectly normal to feel anxious about feeling inadequate. Trust me, you’re in good company. Some of the smartest people in the world haven’t the foggiest idea how AI works.

For a majority of my (meagre) audience, the question is more likely: How is this going to affect my children? Most of us have crossed over to the other side and have either retired or don’t need to hold a job for financial reasons. Our concerns are more about the next generation.

Here is what I urge you to consider: LLMs are not reasoning engines. They are probability engines. They’re not natively intelligent. They are very good at building patterns based on probabilities. If you ask an LLM, “Should I be anxious about AI?” - it will predict the answer you’re looking for and give it to you, while mimicking an expert. The key part is “giving you what you want.” Sycophancy is a bigger problem with ChatGPT, Gemini et al than even hallucination, where the AI just makes stuff up.

But there are areas where its predictive ability is useful. This is where AI becomes the most potent. Writing code is a well structured activity. When you write code, and you write lines A, B and C, there is a fairly high likelihood D must follow, and what D should look like in a given context. Ergo, AI will become better at coding and eventually replace software programmers. I’m sure you can think of other structured activities where AI can be put to use. The thing is - even the most structured jobs have an element of judgement and unpredictability to them. So humans will stay in the picture. Their numbers will shrink but they will be needed.

This - AI cannot reason but only predict what’s next - gives me reason to believe that society might not collapse wholesale. Will AI ever be able to reason? I seriously doubt it. Will AI disrupt the status quo? Yes, in a lot of ways. Even if AI eliminates 20-30 percent of all jobs, it will be cataclysmic. Can your kid avoid being one of those who is disrupted by AI? I don’t know if it’s possible to fool proof oneself from the coming tidal wave, other than learning to swim. A learning mindset will be super critical. Learning how to learn, the ability to self-teach, and a solid work ethic will never go out of fashion. I’d argue - encouraging your kids to teach themselves something creative and fun like a new language or an instrument - will not only enrich their lives but add a super critical life-skill to their repertoire. They will learn to teach themselves and that will pay off over a lifetime.

---

Speaking of historic transformations, the biggest in human history may have been the evolution of languages. Not programming languages but human ones. Around the time humans began to farm, there were about ten million people on Earth. They spoke about ten thousand languages. No language had more than two thousand speakers. Today, eight billion people speak about seven thousand languages, which fall into 140 families. Most of us speak languages that fall into just five of them. Of these, two families are behemoths: Indo-European, whose representative is English, and Sino-Tibetan, which includes Mandarin Chinese. If you include second or subsequent language-speakers, Indo-European is by far the largest language family the world has ever known. Almost every second person on Earth speaks Indo-European. Where did these Indo-European languages start? That is the subject of the book I’ve been reading these last couple of weeks. The book is titled “Proto,” the name assigned to the mother of Indo-European languages. If this interests you, check it out.

---

I have only vague memories of the weeks leading up to the day I became a US citizen in 2002. But I remember the ceremony itself vividly. It was at the Masonic Center in San Francisco. I felt moved by it. It was startling because I hadn’t expected to be moved by it. It was supposed to be purely transactional. We only got our US passports because we were moving to Singapore (and later India) at that time.

I was born in India. Giving up my Indian citizenship felt like a betrayal, a fracture in my identity. India is the mother who raised me, who formed who I was until I turned 21. But America is the mother who adopted me, who opened her arms and showered abundant love and opportunity on a brown kid who showed up on her shores from ten thousand miles away. I am the child of two beautiful mothers. That is how it will be until the day I die. I’d dare you to find a luckier man than I am.

I feel the weight of that destiny more than ever this week as America turns 250. This country is by no means perfect. She carries a deeply flawed, heavy history written in the scars of slavery, racism, patriarchy, and exclusion. It is a long, painful list. Even so, the light outshines the dark. America has, at her core, been a monumental force for good in this world. And the best part of this great country? Its people. I have never met people with bigger hearts, or a more restless desire to do better, than Americans.

That is why I am proud to be an American. May God bless her. Happy 250th birthday, my dear.

Have an amazing weekend!