We no longer need everyone to publish their information in a machine-readable form, because through LLMs we can make machines read the human-readable versions now.
LLMs are not quite reliable, but the machine-readable data was never reliable either. The Web as a whole is messy, and invisible metadata is systematically less well maintained than the primary human-centric information.
besides web, there are tons of data providers and their clients, who are interested in well maintained quality of data..
That's Semantic Besides, not Semantic Web then.