MFA Computational Arts Blog Week 18

Here’s my notes again


Computability, Incomputability and Positive Errors

What does it mean to think of intelligence not as something that can be described completely, what is it to consider it open rather than close. What does it mean to consider the recursivity of algorithms?

Ian Cheng

  • Seeing how agents playing by their own rules within a space, a balance and intersection can be found.
  • Society of minds inside one’s head
  • Abandon a singular model of intelligence. Lets play all models in the same game, rather than seeking a closed definition of intelligence.
  • Avoid a fixed static context of intelligence


  • Think of something, whether machine or brain, is open to the environment, hits, errors, traumas, see how these external inputs have a role in changing the structure
  • The positive definition of error
  • Glitch is that which stands outside the hegemonic interpretation of things, the error to be corrected or cute enough to be considered keepable, but the glitch is abnormal to the mainstream “correct” ontologically legitimate expression
  • Second Order cybernetics
  • Emergence as constitutive of new forms of identity
  • Intelligences need moments of error in order to exist
  • The function of the brain is not the representation of but the adaptation to the external environment
  • Cognition incorporates all external function
  • Technologies of augmented intelligence, adjustment to the external environment often involves destroying the previous cognitive environment
  • These injuries demand the intelligence adjust itself to the new environment
  • Intelligence is always artificial because there is no natural intelligence created that remains static, it is a process of augmentation
  • The more a process implements what it encounters, the more complex it becomes, and the more it is shaped by that environment.


  • Is the critique of teleological reason, of automatising everything towards a specific goal: production control, wealth, power, is this critique still valid?
  • This critique is rested upon a binary opposition: the human and the automation of a process
  • We can replace the human and the machine as the incomputable and the computable.
  • The question may be reasonable for the industrial age, but automation now is a process that adjusts itself based on the thing it is processing.
  • The output of the process readjusts to the process itself, from a closed, first order cybernetics to a second order cybernetics where the question of how to deal with an encounter is up to the processor.
  • The randomness of Incomputability is now intrinsic to the processes of computation
  • On one hand, contemporary capitalism is integrated with second order cybernetics, on the other it is challenged by it
  • When feeding back into the algorithm, the output is therefore larger than the input and cannot be contained/fully understood within the inputs.
  • Incomputability is no longer outside the space, but is intrinsic to computing
  • Alien mode of technological thought that escapes any stable organised finality
  • The technologies for production and control are both vulnerable to changes that destroy their definition and this is also their strength due to being the process of adaptation.

Catherine Malabou

  • Plasticity, the brain as plastic. Literally, when materially wounded, function changes
  • Plasticity changes the idea of trauma psychoanalytically, from being the origin of ourselves, but as things that interfere with the running of the brain and actively, positively create new forms of organisation
  • The brain is in a state of continuous active trauma.
  • Ted Chiang’s “Understanding”
    • The drug only augments the brain when it has been wounded
  • If something increases its ability to function….
  • The brain is not simply wired to everything, it is plastic it adjusts itself to the very world it is part of.
  • The brain is malleable but actively changes itself


            We do not have Artificial Intelligence today, but we do have other stuff like
computer vision systems, robotic abilities to move around, gripper sys-
tems. We have bits and pieces of the grand idea, but those pieces are big
industries. They do not fit together to form one super thing. Siri can talk,
but she cannot grip things. There are machines that grip and manipulate,
but they do not talk. […] There will not be a Singularity. (Sterling 2015)

About this point with regards to industries and fitting

The crowd intelligence – splitting the halves of the brain and asking people to write things produces different results

Roger Sperry split brain experiments

The algorithm that has reached the halting point of computation is much longer than the algorithm that describes the axioms and rules

What is trauma, a moment of rupture for a machine?

Trauma as figurative, the algorithm has to be able to symptomatize and learn to avoid crashing

Strange situations where we can no longer make the differentiation between the machine and the organic because the machine can learn and adapt.

Imagery related to artificial intelligence  suggests that it can be designed, something that can be designed is finished and thus the engineering side cannot guarantee the complete design if the intelligence is to augment.

In the spaces of randomness and incomputability is the point where artificial intelligence does things.

Pasquinelli uses flat ontology in the sense that input = output. The initial description is all that is necessary and sufficient to determine the effect. Cognition is the process of augmentation of the initial condition

Ian Cheng’s simulations allow the machine to work with a moment of randomness.

The web is more likely to be intelligent than the self-contained algorithm

The traditional critique of the socio-political structure no longer applies, we can no longer use critique of instrumental reason – goal-oriented control. The goal keeps changing and the system of control morphs around us and itself faster than we can escape it.

We make our brain but don’t control it. (Marx – we make history but do not control it)

The brain is not genetically deterministic, the brain can collapse its organisation.

Science and technology produces computation and computation proves it is incomputable

The notion of intelligence is constantly tangent to everything we do on the course. It goes hand in hand with the idea that it has to be plastic, it cannot be designed in full beforehand.

Information entropy, the more things happen, the more the experience of those things become part of the mind. In order to describe what the algorithm is going to do, one cannot simply use the starting definition.

Human labour being reframed to repeat the mechanistic production.

Categories: Uncategorized

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s