Strange Entities V: Algorithms — The Monster That Has No Mind

Algorithms increasingly inhabit everyday life.

They recommend films.

Sort information.

Guide navigation.

Evaluate applications.

Generate text.

Shape financial systems.

Influence social interactions.

They seem almost alive.

People increasingly speak as though algorithms possess intentions:

"The algorithm wants engagement."

"The algorithm decided."

"The algorithm knows what I like."

Yet a peculiar question emerges:

Where exactly is the algorithm?

The answer initially appears simple.

One points to computer code.

Mathematical procedures.

Software systems.

Machines.

But difficulties quickly emerge.

The same algorithm can run on different hardware.

The same code can behave differently in different environments.

The code itself may change continually.

The data change.

The interactions change.

The outputs change.

And yet people continue speaking of:

the same algorithm.

The object begins slipping away once again.

The object trap

Object-thinking reaches for its familiar response.

Perhaps the algorithm simply is:

• the code
• the software
• the machine
• the instructions
• the mathematical procedure

Yet every candidate quickly becomes unstable.

Code alone does nothing.

Machines alone do nothing.

Data alone do nothing.

Instructions alone do nothing.

The supposedly self-contained object again begins dissolving.

The monster appears

Now the algorithm starts behaving strangely.

It appears capable of adapting.

Learning.

Predicting.

Influencing choices.

Generating unexpected outcomes.

Sometimes even its creators cannot entirely predict its behaviour.

Worse still, people increasingly attribute agency to it.

The monster appears to possess intelligence without a clear mind.

It shapes activity without possessing obvious intentions.

It acts without an identifiable self.

Object-thinking begins becoming uncomfortable.

Where exactly is the thinker inside the machine?

The relational turn

Suppose the difficulty once again lies elsewhere.

Suppose the puzzle begins with assuming that agency and intelligence must belong to self-contained entities.

Then something shifts.

The algorithm no longer appears as a hidden mind inside software.

Instead it becomes visible as an ongoing organisation of relations.

Code.

Data.

Hardware.

Users.

Practices.

Objectives.

Institutions.

Feedback processes.

Symbolic systems.

None alone constitutes the algorithm.

Yet through their continuing organisation, relatively stable patterns emerge.

The algorithm exists.

But it exists differently.

Not beneath activity.

Within organised activity.

The revelation

And now another hidden assumption begins revealing itself.

Object-thinking often imagines intelligence as a thing contained inside an individual mind.

But algorithms complicate this expectation.

Because what appears as intelligence increasingly begins looking like an emergent organisation of relations.

The algorithm was not the strange entity.

The strange assumption was:

if something behaves intelligently, there must be a mind hiding inside it.

The monster reveals another hidden expectation.

And one final creature waits deeper in the cave.

The oldest monster of all.