I study how language models reason and where they break — that is my day job at work. But the thing that has been on my mind lately did not come from any of that. It came from a story a colleague told me, and I have not been able to stop thinking about it since.
He was at a get-together in Bangalore, with crowd from— a few startup people, a teacher, someone who runs a kirana store near his place, a retired uncle who mostly just listened. At some point one of the founders pulled out their laptop and asked the kirana guy -- what was their biggest problem at store. And the man said: “Kab kya khatam hone wala hai, pata hi nahi chalta.”
So the founder started talking to “Claude Code”, explaining what he wanted in plain language, going back and forth, the way you would explain something to a friend who happens to be very fast at building things. When the output was not right he said fix this part, try that instead. Some people call this vibe coding, I am not sure the term has fully settled yet. About five minutes later he had a working prototype that tracks inventory and sends a WhatsApp alert when stock gets low. I keep thinking about the economics of it — a software agency in Bangalore would charge eighty thousand rupees and take six weeks for something like that. Maybe more.
The kirana owner watched the whole thing and said: “Yeh toh hum bhi kar sakte the.” We could have done this too. And he was right, except for the tense — he could not have done it last year. He can do it now.
I have seen plenty of demos and prototypes. But watching this happen in someone's living room, in five minutes, by a person who was just talking to a screen — I don't know, that felt different. Not like a better tool. More like people who could never build software before can now just describe what they want and get it. Which made me wonder — has something like this happened before? Not just technology getting better, but a moment where building things suddenly became accessible to people who could not do it before?
That is when I recalled Boris (he created Claude Code) mentioning an analogy to the printing press on a podcast. I had not thought much of it then, but now it made me curious. So I started reading about Gutenberg, and the parallel is harder to ignore the more you look at it.
Before the printing press, most people could actually read — that is not where the problem was. The bottleneck was production. Who got to make a book. That was controlled by scribes and scholars who had the training, the tools, and the institutional backing. If you wanted to produce something written and distribute it, you basically needed their permission. There was no way around it.
Gutenberg changed who could produce, not who could consume. Once books could be printed cheaply, the scribes lost their gatekeeping role. And then things happened that nobody expected. There was this German monk, Martin Luther, who had been criticizing the Catholic Church — but honestly, others had been saying similar things for decades. What was different this time was that Luther could print his arguments and get them out to thousands of people without going through anyone. That is basically what kicked off the Reformation.
I think software has been in the same situation for about seventy years now. If you wanted to build a tool or an app or any kind of system, you needed to know how to code. Or you needed to find someone who did, explain what you wanted, pay them, and hope nothing got lost in translation — which it usually did, at least partially. The kirana owner I was just talking about understood his inventory problem better than any developer in Bangalore or San Francisco ever could. He lives inside it every day. But he could not build anything because he did not know programming. That was just how it was.
That barrier is gone, and I think what matters most is not what happens to the software industry — though that is changing too — but what gets built now by people who could never build before. In India especially, and I say this as someone who has lived here and worked here, so much of how the economy actually works depends on very local knowledge. How a particular supply chain in a specific mandi operates, how a panchayat office in UP tracks services, how a school in a small town in Rajasthan manages which students are falling behind. A developer sitting in an office somewhere else was never going to understand that context well enough. The person living inside that problem can now build for it directly.
Gutenberg was trying to print Bibles faster — that was literally his business plan. He had no idea his press would end up reshaping who has political authority in Europe, or that it would help trigger the Scientific Revolution, or that it would eventually lead to the idea that everyone should be able to read. The biggest consequences of what he built had nothing to do with books. They showed up in places nobody was thinking about.
I think we are probably at a similar moment right now and most of us do not fully see it yet. Developer roles are already changing, small teams are building things that used to require fifty people, and people with no technical background are making their own tools. That part is visible. But if the Gutenberg parallel means anything — the changes that will matter most are ones we cannot see yet. They will probably show up in healthcare or agriculture or local governance or education, in some district or town where someone finally has a way to fix the thing they have been complaining about for years.
That kirana owner was not impressed by the demo, by the way. He was not amazed. He already knew exactly what he needed built. He just did not have a way to do it until now.