> complete New York Times (pre-1930)

https://archive.org/search?query=title%3ANew+York+Times&sort...

> as a full PDF download set

I imagine it's possible to achieve this through torrents from Anna's, but you'd have to search and compile the list of all individual PDFs.

> something new with an AI-powered search

With enough time and willingness, someone could put all the old NYT issues through optical character recognition and convert them to text; then make it available to large language models for semantic search of some kind. Ideally public cultural funds could support the effort as academic research.