My batch Whatsapp group has been super active since the last couple of days (above is a group photo taken at the 10 year reunion in 2017).
AI<>Future was a recent topic that attracted many POVs. I am attempting to summarise the overall thesis that emerged. Feel free to drop your personal POV in the comment.
Firstly, there was a lot of disagreement about where this ends:
Some said India is resilient because we’re agrarian and families “take care” of each other.
Then there was a darker view (including mine) - that the government and industrialists might just collude to create a “suppressed regime” to keep the 90% occupied with random things.
Some also seemed to think it could lead to civil war if the educated unemployment gets too high.
But when the middle class is too divided by hierarchical identities, would they ever mount a common cause if they get disrupted?
Personally, I think those who manage to be in the top percentage of people who have an AI agent with enough leverage that the top industrialists / government can’t / won’t own entirely - they might have a justifiable role in the future.
For everyone else, we’ll just have to hope the government does the right thing. Can’t really imagine the current political / economic fabric lasting through this at all.
What all of us agreed upon was that we’re clearly heading into a “jagged transition” where AI isn’t just coming for the boring, mundane tasks, but for the “reasonably intelligent” jobs we thought were safe.
One batchmate mentioned he doesn’t even create his own emails or presentations anymore - his agent does it.
Another batchmate said the most important skill is changing from: “knowing how to do the work” → “knowing what work is worth doing”.
In a world where AI is commoditized, the only real differentiator left could be the human talent guiding those models. Though you wonder - how many humans do you actually need for that? A slimmer human workforce working in conjunction with AI seems to be the likely path. But that leaves many questions for those who will not get to be part of this “human-guide” workforce.
Also, if code and analysis can be offshored to tokens instead of people, where does that leave all our service-oriented tech companies? More so if the gains from AI only stay with a few hyper-scalers in the US, that’s another problem because those profits aren’t going to redistribute to us here. We’d just be providing the data and getting none of the wealth.
So for a country like India, either we find a way to own our own “sovereign” technology stack, or we end up in a situation where the middle class loses its upward mobility and things get... well, unstable.
The irony
There was this one anecdote discussed in the group chat about auto drivers in Bangalore wearing GoPros. They are being paid to provide cheap training data for robotics. It’s a bit chilling when you think about it. They are literally being paid to train the models that will eventually replace them, and they probably have no idea.
Crazy times.
If you are a new subscriber, here’s my short bio. Think of this newsletter as a means to be in touch with a curious friend who is still figuring out life. And if you are in the same boat, all the best.