When I posted the last blog here, vibe coding hadn't yet taken off to the unreal degree it has today (or at least I didn't know about it). Even then, I thought it was interesting that I've thought about the higher and higher barriers of entry when it comes to MVPs and general software development, where if one have specialised technical knowledge and instead decided to focus on other aspects of a business, one finds oneself increasingly "priced out" of the technology ecosystem.
I recently listened to Stephen West on Philosophize This! and came across a technocracy, and realised that this is exactly what I had been pondering (turns out I don't have an original thought anymore). Technocracy essentially meaning a society wherein relevance is determined not by the public or even a government or investors, but by smaller group of technical experts. The example that Stephen gave was that of Amazon, wherein "some guy called Jeff" determines what you see and (to some degree) what you buy, and on the seller side, determines who gets a sale, and is thus able to maximise Amazon's margins. This level of technical expertise determining consumer purchases creates a situation where unless you are able to build the most sophisticated algorithms and capture market share, you're increasingly just a cog in the system rather than one adding to the system.
When I first heard about vibe coding, I was hopeful. Perhaps this could be an equaliser, where technical knowledge would no longer be a blocker for one to validate an idea. I pondered on it for a bit and hoped that vibe coding could be the answer to technocratic systems. It makes sense from one angle. AI can do a lot. I read Sherry Jiang's Twitter where she helped folks ship products in a single day via vibe coding.
But over time, I've started to question whether it actually solves the problem. It definitely helps people move faster at the start, but it doesn't really remove the need for technical depth. It just pushes it further down the line. You can build something that works on the surface, but the moment you want to change how it behaves, or fix something that breaks, or make it scale, you're still going to need proper engineering support. So while vibe coding is helpful in getting something off the ground, it doesn't fundamentally shift who has control or how things are built.
The tools make things feel more accessible, but they also hide a lot of complexity. That's not necessarily a bad thing, but it can create a false sense of independence. You're not always sure what's going on underneath, which makes it harder to troubleshoot or improve things later. It's useful for testing ideas quickly, and there's real value in that, but it doesn't remove the structural gap between technical and non-technical people. It just makes it easier to ignore for a while.