Workshop on "Objects of Interest and Necessity"

For workshop Friday Sept 19, 2025 and Saturday Sept 20, 2025 (with Code&Share) at the Royal Danish Library.

https://aulkalender.kb.dk/event/4407115

https://billetto.dk/e/code-share-59-autonomous-ai-imaging-objects-of-interest-and-necessity-billetter-1610848

INTRO:

What is an "object of interest"?

"Kepler Object of Interest": Potential exoplanets in the Milkyway

 During my eight years of going to the Keskins’ for supper, I was able to squirrel away 4,213 of Füsun’s cigarette butts. Each one of these had touched her rosy lips and entered her mouth, some even touching her tongue and becoming moist, as I would discover when I put my finger on the filter soon after she had stubbed the cigarette out.The stubs, reddened by her lovely lipstick, bore the unique impress of her lips at some moment whose memory was laden with anguish or bliss, making these stubs artifacts of a singular intimacy. 
Orhan Pamuk, The Museum of Innocence,

4,213 cigarette stubs, many with lipstick stains, displayed on needles with white background. Created using Draw Things with the diffusion model SDXL Base (v1.0).

Building an imaginary (in a Wiki)

What is 'autonomy'? What does it mean to separate from capital interest?

Not 'one thing' The range of agents, dependencies, flows of capital, and so on, can be difficult to comprehend and is in constant flux.

This, we have tried to capture in our description of the objects, guided by a set of questions that we address directly or indirectly in the different entries to our 'tour guide'

1. What is the network that sustains this object?
2. How does it evolve through time?
3. How does it create value? Or decrease / affect value?
4. What is its place/role in techno cultural strategies?
5. How does it relate to autonomous infrastructure?

Making maps

"The map is not the territory" (Alfred Korzybski)

Maps are also what 'makes' AI a reality

The corporate landscape of Big Data in 2012 (by Matt Turck and Shivon Zilis).

A map of AI as an instrument of knowledge (by Vladan Joler and Matteo Pasquinelli).

Numbers of newly funded AI startups per country (by Visual Capitalist).

Map of generative AI (by Taller Estampa).

INSIGHT: AI reconfigures and intersects with many different realities.

It exists on different 'planes'

Agenda

Visiting our maps - the different 'lands' of AI imaging - with practical experimentation on what to do in these lands

1. The Land of Pixels (Pixel Space) // 30 mins, or so - installing software
2. The Land of Latents (Latent Space) // 90 mins, or so - making 'LoRAs'
3. The Land of GPUs (Materials/Organisations) // The different 'planes' of AI // 45 mins, or so - making images using your own computer or Stable Horde (P2P), and spending 'kudos'.

PART ONE: 'The Land of Pixels'

A map of AI image generation separating 'pixel space' from 'latent space', but particularly emphasising the objects of pixel space, operated by the users. Pixel space is the home of both conventional visual culture and a more specialised visual culture. Conventionally, image generation will involve simple 'prompt' interfaces, and models will be built on accessible image, scraped from archives on the Internet, for instance. The specialised visual culture takes advantage of the openness of Stable Diffusion to, for instance, generate specific manga or gaming images with advanced settings and parameters. Often, users build and share their own models, too, so-called ''LoRAs'. (by Christian Ulrik Andersen, Nicolas Malevé, and Pablo Velasco)

PART TWO: 'The Land of Latents'

The map reflects the separation of pixel space from latent space; i.e., what is seen by users from the more abstract space of models and computation, i.e., latent space. It particularly emphasises the objects involved in model training (the stacking of latent spaces), but also how latent space is dependent on and array of organisational, technical, textual and visual resources (by Christian Ulrik Andersen, Nicolas Malevé, and Pablo Velasco)

Examples of LoRAs (making stories and 'fixing' things)

LoRA "The Incredible Hulk (2008)."[1]

• 
  An image of Louise Bourgeois with the Real Vision model
• 
  Louise Bourgeois (Real Vision)
• 
  A screenshot of a search query for Louise Bourgeois
• 
  Selected images from the search results
• 
  Annotations for the dataset in the Draw Thing interface
• 
  An image generated by Real Vision with LoRA
• 
  An image generated by Real Vision with LoRA

Men doing dishes

Women fixing airplanes

Practical example: make a LoRA with Draw Things

1. Open Draw Things -> Go to 'PEFT'
2. Select a "Base Model" (to fine-tune) // maybe install a 'light' one first to speed up things
3. Name your LoRA & Select a 'trigger word'
4. Define 'Training Steps' (scroll down to find), set til e.g., 1,000 // the more, the longer the wait
5. Add images to train on. 12 images will work. We have prepared a folder as an option.
6. Annotate the images // what to think about - when annotating?
7. Start training ..... WAIT and WAIT and WAIT (for more than an hour)

Discussion while 'baking' the LoRA

For example:

• How to think of images?
• How to work conceptually?
• How to annotate and what to think of?
• The relation between prompts and models?
• How to make pipelines of models/dataset?
• ?

PART THREE: 'The Land of GPUs (Materials/Organisations) ' ... on the different 'planes' of AI

A map of how objects not only relate pixel space to latent space, but how they are always suspended between different planes - not only a technical one, but also an organisational one, a material one, and potentially many others (capital, labour, knowledge, governance, etc.) (by Christian Ulrik Andersen, Nicolas Malevé, and Pablo Velasco)

The map reflects the material organisation, or infrastructure, of AI image generation. This particularly concerns the use of GPU and the processing power needed to generate images and train models and LoRAs. In conventional visual culture, users will access cloud-based services (like OpenAI's DALL-E, Adobe Firefly or Microsoft Image Creator) to generate images. In this client-server relation, users do not know where the service is (in the 'cloud'). In autonomous AI visual culture, users benefit from each others' GPU's in Stable/AI Horde's peer-to-peer network - exchanging GPU for the currency 'Kudos'. Knowing the location of the GPU is central. Users also train models, for instance on Hugging Face. Here, the infrastructure resembles more that of a platform (by Christian Ulrik Andersen, Nicolas Malevé, and Pablo Velasco).