Mastodon Aaron Tay's Musings about librarianship : Do search+large language models or "retriever augmented" models (e.g. Bing Chat, Perplexity, Elicit.org) really work? The evidence so far