What do LLMs have to offer digital research? To date much of the attention to research with LLMs has been in the service of supporting or even supplanting existing methodological instrumentaria. In earlier discourse, these substitutive means would have been called the virtualisation of methods, referring to their emulation in a new realm. Whether working with LLMs, chatbots or agents, the question has been how (well) they can do our analytical work the way we usually do it. There are core analytical tasks AI is asked to perform. Can they annotate and classify data sets, extracting overt as well as latent meanings, narratives and discourses? Can they affirm or find fault with findings made by more standard qualitative and quantitative means? Can they synthesise data to a degree that it more or less mimics real-world material? Can they produce reference lists on a subject matter? But there is also some work that strives to repurpose their intended use, plumbing the vector space for social structures and cultural trends. For example, can their biases be studied as societal indicators? Can they show extant and emerging social and cultural associations? The Winter School is dedicated native AI methods and their creative uses.