There's some truth in this piece. We never quite managed to really have a semantic web because knowledge engineering is actually hard... and we publish mostly unstructured or badly structured data. LLMs are thus used as a brute force attempt at layering some temporary and partial structure on top of otherwise unstructured data. They're not really up to the task of course but it gives us a glimpse into what could have been.

Interesting view about the LSP specification, where it shines, and where it falls short.

With all those bots and scripts crawling the Web, some of the semantic web vision got silently implemented.

This would indeed be a nice path forward for HTML. It's much too dominated by JavaScript for now, having standardized semantic extensibility would be just better.