Scraping of websites like HN/Reddit/Lobsters/..
Everything described in this issue.
It should at some interval scrape all the websites + use similar techniques to Google/Yandex to index the web with crawlers.
Combining it with what Perplexity is doing to fine tune GPTs with real time data access.