Would a mind that fully understands everything about itself, with no uncertainty or mystery about its own functioning, still be capable of creativity and generating novel ideas?

*I can think, therefore the reach of my understanding is incomplete. >< The reach of my understanding is incomplete, therefore I can think.*

Our very ability to think and create is constituted by the gaps in our understanding. This paradox is an urgent one to understand as AI systems evolve increasingly fast. My point is the following: If we understand the limits of our self-understanding in particular, and that of complex systems in general, to be constitutive of creativity rather than a mystery to be solved, we can leverage these inherent limits to better harness the potential of both.

I’ve been thinking about the ideal conditions for ideas to emerge for years. My company Sane focuses on building tools that facilitate creative thinking and the emergence of ideas. I founded the company with Ida, believing technology had immense potential in unlocking creativity — not by automating it, but by enhancing the already present creative capacities of the human mind. Since we started, AI systems have made massive leaps. The early adoption of these systems, mainly aimed at automating existing tasks for productivity, has made it clear that we need to shift focus from productivity in the conventional sense towards creativity. We should focus on building the muscle that let's us use technology to do things *differently*, not just faster.

Beyond my work in this area, I’m an ‘enthusiastic amateur’ of psychoanalytic theory. In short, Lacanian analysis aims to uncover the unconscious structures underlying our conscious experience, often unwillingly expressed in patterns of language and behavior. Lacanian theory, albeit dogmatic and admittedly with its faults, has been the single most impactful framework in shaping my thinking around subjectivity, creativity, and the human experience in general.

# What is the unconscious and how does it interact with the conscious mind?

*The unconscious is structured like a language — Lacan*

Lacan, building on Freud’s work, saw the unconscious as a dynamic system that operates through mechanisms like condensation and displacement, mimicking metaphor and metonymy in language. This language-like system constantly influences our conscious experience. Slavoj Zizek, building in his turn on Lacan, talks about the unconscious as a site of the ‘real’ — that which resists symbolization. It’s all the (language-like) knowledge that affects us but escapes articulation or being fully understood by our conscious mind.

In cognitive neuroscience, the same system is defined as implicit and explicit processes. Our conscious mind interprets and narrativises behaviors and thoughts that originate in unconscious processes.

Depending on the school of thought, the unconscious consists of repressed desires, sensory and affective information from before the threshold of awareness, internalized cultural archetypes, forgotten memories, and implicit habits we cannot explicitly articulate.

# If our behavior is shaped by the unconscious, what role does the unconscious play in our creativity?

If the unconscious is structured like a language, how does creativity emerge from the unconscious?

Freud saw the interplay between the unconscious and conscious mind as sublimation. The creative act is fueled by repressed desires in our unconscious that get channeled into socially acceptable forms of expression. Lacan, focused on the linguistic aspects of the process, saw creativity as a dialogue between the symbolic (language and culture) and the ‘real’, or that which cannot be articulated. The unconscious generates associations and combinations of symbols that break into consciousness as creative insight through metonymy and metaphor.

In ‘In Search of Lost Time’ Proust describes the experience of such a moment in perhaps the most famous account of a creative eruption in literary history:

The narrator, feeling gloomy, is offered a cup of tea and a madeleine by his mother. As he takes a sip with a crumb of the cake, he experiences intense joy, that he realizes is connected to the tea and cake, but that it’s much more significant than just the combination of flavors. Eventually the taste triggers an involuntary, forgotten memory of his childhood when his aunt would offer him madeleines in a flower-filled garden. The sensory experience unlocks a flood of imagery, carrying unexpected emotional charge, allowing access to an inner world that transcends time and space.

Proust’s madeleine describes how creative insight emerges through an association of sensory experience with unconscious memories, but any association containing ‘overlapping’ or otherwise connected symbols in our psyche can trigger a similar process.

# If some creativity emerges from the unconscious mind, how can we leverage it to augment creativity?

How can we consciously create the conditions for creative outbursts if we cannot articulate the unconscious associations before they surface?

Some common strategies could be engaging in activities that occupy the conscious mind, while leaving space for unconscious processing. Walking is a popular tool for thinking. Ideas often emerge during mundane tasks: in the shower, while washing the dishes or peeling potatoes (my personal favorite). Mindfulness and meditation, or the exposure to different sensory experiences create space for the unconscious to do its thing. For years, I’ve written down my dreams in my Apple notes, often still half asleep, to discover them later and gain deeper access to some of the imagery bordering my conscious mind during sleep. Play allows unconscious processes to flourish by engaging our mind in free associations that we often bypass when occupied with more goal-oriented activities.

It’s our inability to fully grasp our unconscious that makes it such a potent source of creativity. Making space for the unknown areas of our psyche is a way of consciously stimulating creativity.

Being human is like trying to see your own eye without a mirror — the organ of perception cannot fully perceive itself.

In the fields of science and technology, the prevailing worldview is a logic one; everything can and should be explained empirically. But what if the inability to explain everything is precisely what is driving us to do so?

The concept of a constitutional lack has a rich history in philosophy and mathematics. Thomas Metzinger’s self-model theory of subjectivity argues that the self is an internal model created by the brain, referred to as the Phenomenal Self-Model. Since the model is a functional construct, our ability to introspect is limited to its content, unable to reach the components of its own functioning. We can reflect on our experiences, but cannot access the underlying mechanisms.

Similarly, Douglas Hofstadter sees the self as an emergent property of recursive loops in the brain. In ‘I am a strange loop’, he refers to Gödel’s incompleteness theorems, which state that in any complex system capable of self-reference, there are statements that are true but cannot be proven by the system itself.

Bertrand Russell’s work on set theory illustrates this with a perfect conundrum: Consider a set of all sets. If this set contains itself, it shouldn’t. If it doesn’t, it doesn’t in fact contain all sets. The set must both be and not be a member of the set it contains. Russell’s paradox shows the limits to the existence of fully self-explaining systems.

Returning to psychoanalysis, Slavoj Zizek draws on Lacan to argue that human subjectivity is fundamentally structured around a lack or a void — the subject is inherently empty, “nothing but the failure to become itself”. Attempts to fully realize or understand ourselves are always incomplete. This incompleteness is constitutive of the subject.

If any complex system capable of self-reference has some truths about itself it cannot prove, isn’t such a system defined by its limitations more than the nature of its capabilities? Since we can’t examine ourselves (or anything else for that matter) from outside of ourselves, any attempt to fully understand the self from within is doomed to fail.

If we follow this logic, a system fully capable of explaining itself would no longer be intelligent. If I had full access to my own mind, I would no longer perceive it as such. The lack of distance or directed perspective would result in a kind of short circuit, removing the tension between what thinks and what is thought of. Therefore, could we see the unconscious as not an inconvenience making us do stupid shit, but as a constitutive part of intelligence? The view of the self as a construct around a void seems dramatic, but if we look at the history of mysticism and many religious traditions, it’s a core principle that has followed human thought for thousands of years. If we understand our incompleteness as the prerequisite for creativity — for producing new knowledge — we can tap into it as a source.

# Unconscious drives 

Unconscious drives, in the Freudian sense, are those forces that operate outside of our conscious awareness but greatly influence our behavior and mental processes. To make my case, I want to isolate the concept of drive from the properly human (sensory, embodied, affective), and define it through the following key characteristics:

- Emergent — drives arise from interactions within a system, not explicitly or consciously intended.
- Persistent — drives influence us over time. They’re resistant to immediate change. You can take the man out of X, but you can’t take the X out of man.
- Unconscious — drives function below the threshold of awareness or explicit representation.
- Directional — drives are oriented towards specific outcomes, even if they are not consciously recognized.

The black box nature of deep neural networks makes them operate in inexplicable ways. AI exhibits behaviors that are emergent, persistent, unconscious, and directional. An unprogrammed bias somewhere results in a model adopting a character uninstructed, or making a semantic leap that its human counterpart wouldn’t have made. In this vein, the differing molecular / binary constitution of each system is less interesting to examine than the symbolic, linguistic order they both operate in, with its mechanisms of metaphor and metonymy.

There has been a lot of conversation recently around whether AI will be able to produce net new scientific research without a provided theoretical framework (ref. Mel Andrews). If we follow our line of thought, could we project the ‘drives’ in AI systems, as defined by their key characteristics, to develop into directions similar to the workings of the human unconscious in enabling creativity? With the risk of drawing an overly ambiguous line, can we separate those elements in our unconscious processes that are not dependent on embodied experience, and see similarities in a potentially creative machine unconscious? Can we see intelligence as neither human nor machine, but isolate those parts of it that are apparent in both? It’s common knowledge that the human body consists mostly of water, but we rarely discuss the degree to which the sea resembles us. In a similar vein, we don’t need to anthropomorphize AI to recognize similar principles underlying its behavior. A lot of the same, much of the different.

If we see this lack, gap, or incompleteness in understanding through this lens in both human and machine creativity, how does that change our perception of their interplay? Instead of attempting to model ‘creativity’ in machines, can we stimulate ‘unconscious’ states in these systems similar to the strategies we as humans have to create conditions for creative outbursts? What are the machine equivalents of madeleines and tea — the Proustian triggers for AI?

*Work in progress*



Incompleteness as a feature

In the age of second brains, productivity tools, and character limits, everything around us — if that preposition can be used to talk about the internet — seems to be pushing towards concise communication. Reaching wide audiences requires all-terrain language, and the urgency of the present moment, amplified by chronological feeds, doesn’t allow for much stylistic variety. Efficiency is key — compressing as much information as possible to the least amount of words is the ideal of all communication.

While I do agree that many areas of interaction, like good leadership or effective collaboration are [built on concise explanations](https://twitter.com/kepano/status/1693368019723174315) and easily transmittable ideas, the potential of compressed meaning has its limits. When language is reduced in this high-modernist quest for efficiency and common sense, something of the essence of what’s needed to create truly new ideas is lost.

Think about your favourite book, an individual passage that made you feel something, gave valuable insight, or even changed the way you think about the world. Why did this particular piece of writing touch you?One of my favourite authors is Milan Kundera. I first read 'The Unbearable Lightness of Being' in my early teens. Year after year I keep going back to the book, reading it through a new prism each time as my experiences in life grow and shift focus. One of the many passages I’ve kept close over the years is the following: “The brain appears to possess a special area which we might call poetic memory and which records everything that charms or touches us, that makes our lives beautiful ... Love begins with a metaphor.”

When I’m reading, what makes me stop and think is a sort of ‘resonance’. For an idea to resonate, it needs to do the following: 
- Something in my subjective experience or previous insight connects to the idea in front of me and gives it potency. The connection between two ideas always carries the possibility of a third, new idea.
- It moves me through language. I might find the phrasing particularly harmonious, the choice of words unique or playful, or an individual sentence might make me relate to the author in a way that provokes emotion.

Resonance is what makes an idea stick, rather than serving merely as another piece of information. In ‘Observations on ‘Wild’ Psychoanalysis’ Freud notes that if knowledge alone (of one’s symptom) without affect was enough, his patients could be cured simply by asking them to go to lectures or read books, and that “such measures have as little effects on the symptoms of nervous disease as distributing menu cards in time of famine has on people's hunger." A true, transformative insight requires being experienced through affect.

Freud’s metaphor is very much applicable to the information ecosystem we live in today. While we have immediate access to masses of information, we seem to be regressing in terms of how we relate to it (and perhaps, through that, how we relate to each other). Information in the form we consume it today seems to create more anxiety than resonance. In our quest for efficiency, we risk losing the power of language to deeply affect us — to touch our poetic memory. Without affect, we keep accumulating information to optimise towards goals based on existing structures, rather than striving to create new ones.

Historically, thinkers with a distinct or difficult style have often been celebrated posthumously. Kafka, Hegel, and Proust all rose to significant fame only after their deaths. The impact of rich, ornament language can be less immediate than that of concise writing, due, perhaps, to the smaller initial reach. However, the depth of insight it can give to the reader willing to go through the experience makes for texts that linger in time, and ideas that have the potential of infiltrating culture at large. This isn’t to say that concise text can not provoke affect and insight, but something about prolonged attention and contemplation give playful, even complex writing its disruptive power. It makes the potentially resonant surface bigger.

Common sense, like common language, reflects current reality. By relying solely on it to communicate on ideas we risk losing out on a different one in the future — in order to think in new ways we need new language.The tools we communicate with online should enable unusual, exceptionally creative ways of expression and language to be shared and discovered. There is no reason, if not the limitations of the interfaces we use, for wild ideas on the internet not to have the kind of lingering, evergreen value a classic book has — ideas that resonate on a deep level, that we can revisit, that evolve over time. Even ones that don’t immediately take hold.

Language and Resonance in the Age of Efficiency

A lot of the recent conversation around AI tools revolves around the idea that as the technology advances, we will no longer need specialized interfaces. Instead, we will be able to execute any task with natural language commands. Whether the interface metaphor is a command line, a chatbot, or a search bar, the premise is the same. Input words, output magic.

The pro-chat position assumes that humans are inherently good at communicating their needs. But isn’t communication and understanding THE fundamental human problem? Natural language is by nature ambiguous, context-dependent, and imprecise. There are as many interpretations of a given sentence than there are people — language is burdened with subjective associations, and we are burdened with having to communicate with that language. We are notoriously bad at precisely articulating our needs. Doing so requires us to first effectively identify them, which is a hurdle on its own. Learning to ask the right questions is a quest of a lifetime — it seems suspicious that we’d count on it for the adoption of the most powerful technology yet created. 

The concept of the Freudian slip effectively illustrates my point: In everyday discourse, we often believe we’ve misspoken when we say something unintended. However, from a Freudian perspective, these ‘slips’ are not mistakes at all, but reveal unconscious thoughts or intentions that we’re unable to symbolize and therefore interpret as ‘random’. This is to say that our use of language is anything but a universal application for a fixed syntax of meaning. Replicating human communication as the ‘ideal’ non-interface is denying the ambiguity of human thought and the fluidity of language.

Observing friends and family firmly outside of the tech umbrella try to use ChatGPT has led me to think that there’s fundamental faults in this approach. What often happens is some version of the following:
- The person assumes the AI has more context into their intent than it does. 
- Because of the above, the person doesn’t understand the extent to which the structure of their command affects the AI’s output.
- The frictionless nature of the chat interface paired with the user’s poor prompting eventually leads to the user becoming bored.

My interpretation is that boredom in this case stems from a perceived lack of agency (“this is as good as it gets?” as in, assuming the AI’s output will be similar in quality regardless of the user’s input, and therefore abandoning the effort). 

The conversational nature of the chat leads the user to anthropomorphize the AI, leading to a false sense of comprehension — assuming a level of understanding that doesn’t exist. Paradoxically, relying on natural language as the interface for AI requires us to learn a new language. The linearity of chat presents other issues: it makes it difficult to iterate on results, to branch out from an output, and navigate to or reference previous points of a given process. 

# The interface imperative

Consider a future where any task commonly executed on a computer can be done by voice command, or a universal search bar / command line. The lack of visual cues, processes, and elements to manipulate leads to confusion and boredom. As a reminder, *‘What time is it’* constitutes a surprising amount of Google search queries.

If we adventure further, why not execute tasks by simply thinking of a prompt? The boundary between intent and action becomes blurred. I can now do ‘anything’, but paradoxically, if so, what should I do? Anxiety stems from a perceived excess of choice. I struggle to distinguish between my own thoughts and intentions and the AI's outputs. The erosion of boundaries leads to a a short circuit of action and consequence, and a diminished sense of self and agency. 

# The cognitive load paradox

In open-ended interfaces, like a chat, the perceived cognitive load is heavily on the machine. The user is told ‘ask me anything (and I will do all the work)’. In reality, the effective load falls on the user, who needs to conceptualize their needs without guidance, articulate them precisely, and navigate the obscure, seemingly infinite capabilities of the tool. This misalignment leads to the loss of the user’s sense of agency, boredom, disengagement, and generally disappointing outcomes.

The other extreme are interfaces that place the perceived cognitive load on the user — tools with a lot of specialised features that distract the user from the task at hand and lead them to underutilize the tool’s capabilities. 

In an optimal scenario, the perceived and effective cognitive load is balanced between the user and the machine by the interface. The user feels in control and understands their role in the interaction. Their perceived agency matches the actual influence they have. The interface provides sufficient structure to frame the interaction, and gives real-time visual feedback to the user. 

# Transitional spaces and sufficient frictionless

The (grahical) interface serves a purpose far more profound than its mere functionality. Consider this analogy with the Lacanian concept of transitional spaces: A analyst's office serves as a necessary frame, a controlled environment where the patient is free to self-explore and grow. Within this space, the patient can articulate their experiences, thoughts, and emotions in a way that brings them insight to their psyche. While the patient is certainly free to articulate their experiences outside of sessions, these expressions don't carry the same transformative power. The therapist's office, as a transitional space, is the essential context for analysis to happen. In this vein, in the context of human-computer interaction, the ideal interface sufficiently delimits the potential or suggested course of action for the user, mediates a feedback loop that’s meaningful, and preserves the user’s sense of control over the output. 

# A case for spatial

The human experience is inherently a spatial one. We relate to the world through our body as it moves through and manipulates the surrounding enviroment. Our interpretation of spatial relations between things has a semantics of its own, that we can translate into words, like ‘close to’ ‘under’ ‘along’ or ‘on top of’. However, our understanding of these relations is pre-linguistic — we understand them instinctively. 

None of this is new. The concept of direct manipulation was introduced in 1982, following principles that are just as, if not more relevant in the context of AI: Continuous representation of the object of interest, physical actions over complex syntax, continuous feedback, and reversible, incremental actions. Since, we’ve gotten desktops, touch screens, pinch to zoom. If you’ve ever seen a baby use a touch screen before they can speak, you’ll understand my point. 

It’s underwhelming that with all the power of AI at our fingertips (pun intended) we’re now rebranding the terminal as the accessible interface of the future. Meh.

In the case of AI, the main advantage of a spatial interface is that it allows for visualising complexity without increasing the cognitive load of the user, because we naturally understand spatial relations. It’s not a language we need to learn. We don’t even need to go as far as 3-dimensional space to benefit from this intuition — a 2-dimensional canvas compared to a linear interface already significantly grows the possibilities for manipulating inputs, and visualizing or comparing outputs. 

*Work in progress*

# Materials, not mirrors

To summarize; the goal shouldn’t be making AI more human-like, because humans are inherently confused. Replicating human communication as the ‘ideal’ non-interface is denying the ambiguity of human thought. It’s a lose-lose scenario. A better goal would be to create interfaces that make the capabilities of AI more accessible, comprehensible, and controllable, not more human-like. These interfaces should be flexible, scalable, and modulable, but interfaces nevertheless. 



The Chatbot and the Freudian Slip

In the preface of ‘Spinoza’ by Gilles Deleuze, Robert Hurley writes:

>“I am aware that I have said next to nothing about Spinoza. The fact is that Spinoza is difficult. And this book on Spinoza is difficult. But the situation is helped by the author's word to the wise: one doesn't have to follow every proposition, make every connection — the intuitive or affective reading may be more practical anyway. What if one accepted the invitation — come as you are —and read with a different attitude, which might be more like the way one attends to poetry? Then difficulty would not prevent the flashes of understanding that we anticipate in the poets we love, difficult though they may be. The truly extraordinary thing about Deleuze is precisely the quality of love that his philosophy expresses; it is active in everything he has written. I like very much a phrase in Arne Ness' article, referred to above. Speaking of Spinoza's amor intellectualis Dei, he says that it "implies acts of understanding performed with the maximum perspective possible". As I see it, just such a performance awaits the reader here.”

My encounter with Hurley’s text happened at a time where my mind was newly electrified with the thought of everything there is to read, to learn, to understand — I was trying to ram through both foundational and secondary texts from European philosophers and french theorists’ interpretations of their works. Lacking a formal education in classics or philosophy beyond a few semesters of semiology and Chinese thought (which was the obsession of my art school philosophy teacher), I was often frustrated at the slow process of building a base of concepts solid enough to grasp the authors’ words in one reading.

I had given up on reading fiction altogether. Reading for ‘pleasure’ didn’t seem ‘productive’ enough, it was as if only the mind gymnastics of difficult texts, chosen in order of their historical impact, were worthy and able to occupy the remains of my disappearing attention span that years ago could take on hours and hours of fantasy fiction in a single sitting.

“One doesn’t have to follow every proposition, make every connection — the intuitive or affective reading may be more practical anyway.”

It was Hurley that gave me permission, in a way, to let go of the dichotomy of leisure vs productivity, and to lean into a way of reading that, instead of stopping at every unknown concept or difficult turn of words, made me happily skip over pages, make seemingly unrelated connections, and abandon a book when it no longer seemed to give me pleasurable insight, to perhaps open it again at another time.

Intuitive or affective reading may not be goal-oriented in the way that institutional education requires, or productive, if we see productivity as a quantitative measure of favourable outcomes. But what if productivity, in the context of reading, meant that when there’s pleasure and creativity involved, we might end up thinking in new ways (as opposed to learning to think ‘the same’)? What if a better way to read is through *acts of understanding, performed with the maximum perspective possible?*