Worklab 2024
General info about the Worklab
For years the worklab is a core format in the artistic research program of ‘mur.at’. Worklabs are “un-conferences” that can last for several days, to which people are invited to participate via a public open call. Worklabs create a temporary space in which artists (= specialists in the field of technology, design, art, research etc.) and external guests can discuss topics from the field of digital technology, art and its social implications. People from different fields and with diverse backgrounds meet here and work together on ideas and designs that will become the starting point for future projects, publications and productions of the current and coming years. The worklabs are both incubators for specific work and open think tanks where ideas, tools, as well as practical and creative approaches are developed and tested.
About the Worklab 2024
LLMs represent a major advancement of the ‘computational turn’ in mediality, but remain rather isolated as an object of investigation or critique. This is partly due to marketing hype surrounding them (which becomes the secondary object of critique) or their ambivalent corporate origin, but also a lack of ambitious theoretical considerations. In order to attempt a more wide ranging view we invite you to join this year‘s worklab organized by GIA (General Intelligence Agency of Ljubljana) in collaboration with mur-at community. Together we will consider questions such as:
- How do LLMs remediate old dreams (and nightmares) of universal languages, encyclopedists, archivists and all kinds of information management projects to which LLMs are peculiarly related?
- How is the question of remediation related to abstraction and what kind of vistas do LLMs open for us (new dreams and new nightmares)?
- In what sense can we understand LLMs as the first true AI cultural interfaces?
- What does the future hold in an era of ubiquitous LLMs?
- What kinds of social and self-relations will be afforded by LLMs?
with contributions by: Gabriele de Seta, Sarah Burkhardt, Sarah Fitterer, Sofia Lee, Peli Grietzer, Željko Blaće and others aswell as the GIA collective: Martin Hergouth, Robert Bobnič , Maks Valenčič and Jan Kostanjevec.
Worklab 2024 key dates
- Dates: June 19-22, 2024
- Where: KIG - Kultur in Graz, Lagergasse 98a, 8020 Graz.
All presentations, workshops and discussions are publicly accessible and will be held mainly off-line and in English at No entrance fee.
Location
KIG - Kultur in Graz, Lagergasse 98a, 8020 Graz.
Public meet & greet of contributors
- Planned for Wednesday June 19, 2023, 19:00 at KIG - Kultur in Graz, Lagergasse 98a, 8020 Graz
Preliminary Program!
Workshops and public presentations
Thursday: June 20, 2024: 10~17:00 “LLMs in Systems”
The first day of the worklab will explore the relation between LLMs and the ways in which these models are incorporated in contemporary infrastructures. The datafication of all of the system we interact with has reached new levels of importance, especially when nowadays infrastructures are becoming increasingly interactive and mediated through such „generative systems“. These implications will be explored through politico-economic, socio-anthropological and techno-philosophical lens.
- Presentation: Intro by GIA collective
- Martin Hergouth (moderator: Jan Kostanjevec)
- Sarah Burkhardt: ‘Foundation Models are Platform Models: Prompting and the Political Economy of AI’ (moderator: Robert Bobnič)
- Gabriele De Seta: ‘Philosophical zombies, stochastic parrots, masked shoggoths: Making sense of LLM folklore’ (moderator: Maks Valenčič)
- Workshop by Željko Blaće: Feeding the (Wiki)DATA to…
- Residency Workshop: GIA design prototyping & reading group
- Daily reflection
Friday: June 21, 2024: 10~17:00 “LLMs as artefacts”
In the second day we will shift to a more practical and experimental approach to such models to understand their latent representations. Through a psychoanalytic engagement with the “unconscious” of LLMs, we will try to prompt these models as systems with specific understanding of cultural knowledge, knowledge in general and their internal models through which we interact with them as generative systems for multimodal generation.
- Workshop: AVERAGECORE: ‘Psychoanalysis and Anthropology of LLMs’ by Sofia Lee and Sarah Fitterer
- Daily reflection
Saturday: June 22, 2024: 10~17:00 “AI and Knowledge”
On the third day, we will try to further explore the connection between knowledge and LLMs and in what sense these are not simply cultural, but also epistemological artefacts. As so-called “fundational models” they, almost necessarily, lead us to explore fundamental philosophical and technological implication of their success and what kind of ramificiation this will have for the future of man-machine interactions and generative AIs in general.
- Peli Grietzer: ‘Philosophy of LLMs’ (moderator: Martin Hergouth)
- Robert Bobnič: ‘The use of LLMs in recommender systems design’ (moderator: Maks Valenčič)
- Jan Kostanjevec: ‘LLMs and Everything’ (moderator: Martin Hergouth)
- Maks Valenčič: ‘In what ways are LLMs Remediating Culture?’ (moderator: Martin Hergouth)
- Residency Workshop: GIA design prototyping
- Worklab reflection & wrap up
Sunday: June 23, 2024:
- Individual departure of participants - no official program
Cooperations, funders of mur.at and partners
Photo credit: “Graz - Cityscape with Sunset” by Bernd Thaller, CC BY-NC 2.0.