AI and Intellectual Imperialism

I first noticed it with IT specialists and programmers in Russia.

Their success in one field - and the high salaries that came with it - made them believe they knew better than everyone else. Soon, they started teaching others how to do things in completely unrelated areas, often ones they knew little or nothing about. It’s almost like the Dunning–Kruger effect, but not quite: these people are genuinely competent in their own field; they just overgeneralize that competence to other domains.

Now this intellectual imperialism is spilling everywhere, amplified by AI.

Programming and writing code stand apart when it comes to AI. They have none of the copyright and ownership issues found elsewhere, and nowhere else is its usefulness so obvious, adoption so widespread, or the stigma around using it so minimal. In most other fields, AI is still treated with caution or defensiveness; in programming, it is simply another tool. Perhaps that normalization is precisely what creates the illusion that similar gains must exist everywhere else.

All those AI bros have found a hammer that works great for their nails, and now they’re trying to apply it to every other use case imaginable - even the ones that require a more nuanced approach.

It all reminds me of the square hole meme, where every block is forced into the same square hole. The square hole is generative AI and LLMs, and the AI bros keep trying to fit everything into it. Even if it technically works the first time, that doesn’t mean it’s the right hole for other blocks. The same goes for AI and LLMs: they are not optimal for everything, and the chat interface is not always the best way to interact with computers.

And in this analogy, we’re all Alison on the left - watching, increasingly frustrated, as everything goes into the square hole because those guys “know better.”