Jonathan Blow

Well then Jonathan's point is moot. Web developers make web sites.

I think the point that he’s attempting to make is that web developers could make better performing web sites if there was better understanding of hardware and if APIs were used to take advantage of that hardware, i.e. WebGL. I’m not entirely sure though. That’s how I read it.

WebGL is used to make things like 3D Objects. It's pretty cool. But to suggest that devs should use 3D rendering to create a page with text and images is silly. Even if it would perform better.

Thats the crazy thing though. I promise you it would be less dev time, less buggy, and redonkilously faster using webGL. Because then you actually have the ability to program the gpu. Native gui toolkits like GTK or qt use 3d api's like Vulkan or GL for graphics.

A big part if the internet is the ability to search for and find pages of text. Rendering everything as graphics is a fundamentally bad idea.

What do you think happens when you draw text? How does the text become pixels on the screen? It has to happen somehow.

Yes but if you go full WebGL right now, you lose the plaintext that can be indexed, searched, read by screen readers and so on. You just have an opaque <canvas>.