Altai: An interview with Lukáš Likavčan


For more information on the Alt’ai project, please visit and the Strelka Institute’s Alt’ai subpage.


Vit Van Camp: Am I right in understanding alt’ai as a predominantly aesthetic project, one which broaches themes such as machine-machine interaction and complex adaptive systems, and formally explores the idea of “automated landscapes”? What role do aesthetics play in the overall concept?


Alt’ai: At the beginning, there was a set of research interests – sovereignty without territory, automated landscapes, machine-to-machine interactions, operational images, animist ontology – but in the design process the aesthetic factor became more prominent. In the end, we hope that we have approached the point of fine balance between concept and its formal, aesthetic qualities. Or to put it differently: we have actually never thought of aesthetics as something independent from research interests underlying our project. It would be better to frame aestheticization as part of our design strategy that targets the problem of machine-to-machine interfaces. According to our approach, aesthetics becomes a tool for multiplying the possibilities of future authentication protocols. So in fact, alt’ai has never been that much an aesthetic project; it makes use of aesthetics as an integral component of speculative design.


The aesthetics which the alt’ai presentation evokes provide a sense of the frontier, being almost otherworldly, as if from a distant planet. They are very beautiful. To engage closer with the idea of speculative design, however, it seems that most real-life uses of such a type of automated landscape, one where various machinic agents would interact based on predetermined protocols and in semi-autonomous ways, could in fact be closely forecast in more mechanized environments, such as smart factories, smart cities, freight logistics, etc. What brought you to choose such a rather bucolic world to populate with your machines?


Today, what we mean by automated landscapes are usually industrial parks (e.g. Pearl River Delta), logistic hubs (e.g. Port of Rotterdam) or agriculture fields and pastures (e.g. dairy and tulip farms in Netherlands). Our ambition was to go one step further and to imagine what happens when this trend hits really remote locations, such as pristine valleys and mountain systems. We are already being evacuated from vast landscapes as automation proceeds, so why to stop at the bottom of the mountain? Thus the first reason to choose the Altai mountain region as template for our simulated universe comes from this consideration. Second, Altai is the place of overlapping peoples, cultures and jurisdictions. The amalgam of manifold practices emerging from this mixture was fundamental for our ideation of different rituals the agents in our simulation can partake in. Some of them are religious, like sky burials or pilgrimages. Others are legal, like border crossing. And still another set of practices relates to basic evolutionary strategies (e.g. mimicry) or technical behaviour (e.g. sensing and recognition). Third, we were quite curious about possible co-evolutionary trajectories of autonomous machines and biological species in the wild. What happens when species meet, where by “species” we mean also technical objects with their own developmental vectors which Simondon talks about?


What alt’ai also does is that it “generates special images called cosmograms.” In the whitepaper you understand cosmograms as a type of mediating device, an operational image signifying for both human and machine, but one which potentially communicates with all shades of the man/machine spectrum. Through vision, the images seem to allow on he level of the human observer access into a machine’s subjectivity, while at the same time remaining functional in establishing machine to machine protocols.


I am not sure about the human end of the spectrum of image readability. Yes, our cosmograms resemble human-made pictures, and their visual language comes from research on Buddhist mandalas or Eastern-Orthodox icons. However, on the machinic level, which is the only really important perspective in the context of automated landscapes, the pictorial quality of the image is an accidental surface for essential mediation of algorithmic process (in this case it is authentication), which is the true purpose of the cosmogram. Rather than subjectivity, the insight of the cosmogram lies in the recognition of the entity as belonging to a generic ontological space of an address layer of the stack that the given automated landscape belongs to. In other words, cosmogram is technology that operates according to animist ontology.


Can you elaborate on the concept of ‘authentication’ among the machines, and what role it plays in alt’ai’s overall ecosystem, as you imagine it?


Authentication can be defined as verification of an entity’s identity. Imagine a situation: a drone approaches a checkpoint before entering some warehouse area. It declares to what institution it belongs, what is its flight path and purpose, and subsequently it is allowed to enter an area. How does it happen? Usually, an entity can be authentified according to something it knows, is or has. It is a method of gatekeeping, filtering those entities that can access some platform space or function from those to whom the access is denied. Whenever you pass CAPTCHA test, you go through authentication: authentication procedures many times reenacts the Turing test on daily basis. Our approach generalizes authentication as a procedure that checks some unique imprint or footprint of en entity.


Now comes the funny part. Accidentally, the idea of imprint is used in Eastern Orthodoxy to explain the function of icons: holy images do not represent the person of a saint, they serve as interfaces for religious practices (such as praying to the saint) based on the causal chain that connects the picture to some divine event (a presence of the holy person or some part of his/her body, divine inspiration of the artist), which is then forever imprinted in the image. So what we see here is the medieval use of image as interface, which is so ubiquitous today, plus use of image as a storage of some imprint. This is, by the way, also an approximation of what Peirce had in mind when he was writing about icons in his semiotic theory – an icon is a sign constructed by means of an analogy, and hence it imprints some unique trait of an entity it is derived from.


Following this research, we asked ourselves: can we use images to store imprints of entities for authentication purposes? The answer is obviously positive – just look at how QR codes function today: in a factory we have visited in Shenzhen, each object has a QR code that is its unique authentication device. In this sense, cosmograms are aesthetic radicalization of QR codes for the same purpose.


A question on the workflow in the group and in the New Normal workshops: in the alt’ai whitepaper you write that each participant had to step out of the comfort zone of their respective field of expertise and act in a “deliberately anti-disciplinary” manner. Throughout the project, you essentially had to “become someone else.” How did you personally feel in such a position and what were some of the coping mechanisms you found yourself developing?


Our team was extremely diverse and it included people with very different workflows, which set a high benchmark in mutual tolerance towards not only our skills, but also our personal limits. For example, I [Lukáš Likavčan] spent four weeks programming and designing in Adobe toolkit instead of doing philosophy (which is my original training), but that might have been a bit precarious position, because once you freshly enter a new practice, it is hard to tell what is right and what is not right. Whereas for someone trained in philosophy, it is easy to distinguish bad practice from good, one can get uncertain about programming or design development and its methods at the beginning. The coping mechanism was seeking reassurance in the team, but as the time went, everyone got more and more independant. It seems that the first principle of successful collaboration in such a diverse team is taking each one of your colleagues as smart and autonomous human being. Then the flow of skills and knowledge can freely move in all directions.

Diffract this //

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.