Lately I’ve spent a lot of time wondering where the world is headed, and I’ve not been feeling great.
Let me explain. I’ve spent a huge part of the last few months doing something called “vibe-coding”, I’ve spent more time talking to Claude Code than the people around me, both at work and even outside of work. And the more I do, I realise AI has really gotten smarter than ever before, and it’s only going to get smarter. I think people who are still ranting about AI being stupid are honestly either being delusional for their mental peace or just are using outdated AI tools. I’m not trying to convince those folks otherwise, but if you are one of those folks Matt Schumer’s written a beautiful article about this a few months ago. Give it a read: https://shumer.dev/something-big-is-happening
As somebody who works in a tech-adjacent role sure I’m seeing this probably faster than many other industries are, but this is simply a function of the fact that AI is first being used to build itself, and jobs of it’s builders best. But I can think of very few industries that AI can’t take over. Sure maybe physical labour may take a few more years for the economics of Robotics and AI to meet human wages especially in a country like India, but at the pace at which technology is moving, even that’s not too far. My parents who are both scientists, have started using Claude code at their jobs, and have been mind-blown by the research capabilities and have already written full-fledged publication worthy Academic research.
Well you must be wondering, well all this sounds great, we’re all going to 1000x our productivity this is great. Sure, I hope everybody reading this does, but that’s not really how the real world works. If one person can do the job of 1000, companies will shrink their employee sizes drastically, more people will be unemployed. When people are out of jobs, naturally spending will reduce. Productivity is about making things. But GDP requires someone buying them. If 999 out of 1000 workers are displaced, who has income to purchase the output? You get a glut, not growth. This is the paradox of underconsumption problem that classical economists have warned us of.
Well now you may be asking, but technological advancements have hit us like this SO many times before - the agricultural revolution, industrial revolution, the internet boom and more. We’ve survived all of it, what’s so different this time?
Historical productivity gains were 2-5x over decades. People had time to:
- Retrain for new sectors
- Migrate to new cities
- Build new institutions
This time around the infrastructure to adapt does exist (thanks to availability of internet), sure inequality still lands the best of AI in the hands of the rich who can afford it, and the inequality is only about to widen.
Now you might tell me, “We’ll invent new jobs we can’t imagine.”
This has been true historically. But there’s an implicit assumption, that humans have a comparative advantage in the new tasks. If the thing delivering 1000x productivity is general-purpose AI, it may also be better at the new jobs before humans can retrain for them.
I really wanna be on the side of the optimists who believe the human race has always adapted and found a way to survive, I really do. But day by day, I’m just just getting even less convinced. I was having a conversation with a friend yesterday, who told me “Charu we’re still in the 1% who’s leveraging AI, you’re paranoia is deep because you’re deep in a Bangalore tech bubble. We’re ahead of the curve, this is great for us, atleast for the next couple of years until it catches up to us” And I think she’s probably right, until it catches up to us, we’re safe. For the rest of the world, who doesn’t, they really are running out of time.
PS: Did a small migration and revamp of the site. If you see any glitches or if anything is off do let me know :)