Podcast - https://notebooklm.google.com/notebook/89bac7a8-f0ba-444e-857b-b16400fb014e/audio
In the etymology of the word "gloss," we find Ariadne's luminous thread. Following this thread leads us through a labyrinth of history to our present moment of technological metamorphosis. From the Greek glossa (tongue) through the Latin glossa (requiring explanation), the term 'gloss' evolved to taking on connotations of both addendum and luster. This semantic duality perfectly captures the now paradoxical glittering sheen of the human contribution to artificial intelligence. This applied surface glow, interpretive and important is the glossy sheen that humans now overlay through their experience, intuition and directives. It is also the decisive difference between human and machine. This gloss represents the eternal touch of the master but also classic sculptural beauty and flaws of the "human, all too human gloss". This gloss is the ineffable accumulation of lived experience, intuitive leaps, and the deep strata of unconscious understanding that larger amount of billion parameter AI large language model training have yet to fully replicate. The early 20th century German Jewish art historian and Frankfurt School social critic, Walter Benjamin called this 'the aura' or essence of the original of a painting or work of art. He lamented this aura's loss in the early part of the twentieth century, what he termed the age of mechanical reproduction or what we would now call in the first or second industrial revolution that we have moved wellbeyond. The human gloss though has not yet left our writing completely and perhaps the gloss like co-intelligence will rise anew like a phoenix or the renaissance (rebirth) or at least new partners, our new bedfellow artificial intelligence to remain and blossom.
Consider now glosses' cognates: gleam, glow, glamour, glisten and glitter. Each suggests illumination, but each of a particular kind—not the harsh spotlight of pure logic, but the warm radiance and glow of understanding that comes from embodied experience something also 'je ne says quoi' implicitly glimmering and glittering like an aura but not harshly specified or probabilistically dropped as the final difficult choice. Paradoxically, this is precisely or probabilistically what distinguishes human intelligence from its artificial counterpart, even as the gapnarrows through what has been termed cyborg, centaur or less crudely "cointelligent" connections. This is the emerging synthesis of machine learning's vast probabilistically data driven parameter spaces with humans and their formalized learning and experiential wisdom.
The parallel between neural networks' training data and human experiential learning proves both instructive and incomplete, like the Princeton logician Gödel's incompleteness theorem where no formal system can be both complete and consistent within itself. Just as neural networks, no matter how vast their training data, cannot fully capture the self-referential, meta-cognitive imaginative, intuitive ineffable and imminent aspects of human experience that allow us to reason about the limits of our own understanding. This fundamental incompleteness in both human and machine cognition suggests that their synthesis might transcend the limitations of each, even as it creates new forms of incompleteness at higher levels of abstraction. Also, while modern language models train on hundreds of terabytes of data and text—roughly equivalent to reading millions of books—this differs fundamentally from a human's accumulation of lived experience. A professor with thirty years in their field hasn't merely accumulated facts; they've developed what Michael Polanyi called "tacit knowledge"—the things we know but cannot tell, the intellectual muscle memory, intuition and unconscious perhaps neuronal processes that allow vast creative leaps across seemingly unbridgeable conceptual chasms that take lifetimes of 'multimodal', 'multisensory' experience that comes about from being born an embodied subject to an orchestrated multimedia world.
Consider now the case of Einstein's famous Gedanken (thought) experiments. When he dreamed of an imagined chasing a beam of light, this wasn't merely a processing of physics equations, the processing did follow, but the first was an act or 'day dream' of embodied imagination, drawing on a lifetime of physical and formal physics experience thinking about concepts of motion, light, and space. This purely human example may also be termed "synthetic cognition": the marriage of formal knowledge with experiential understanding, of explicit rules with implicit insight perhaps probabilistic or interpellated but also analogical (creating analogues) and logical, utilizing both logic of formalized physics and its equation to provide new inferred information about matter and energy while keeping the speed of light constant, Einstein's famous, E=MC squared.
The most profound possibilities currently for artificial intelligence also have to do with this gloss and cointelligence of varying systems working together and lie not in replacement but a little displacement and synthesis. Take for example the hypothetical case of a theoretical physicist with additional degrees in computer science, philosophy, and cognitive science who has thrown away her physics position in the academy to pursue a much more lucrative and for now fun career as an online YouTube influencer and pundit. When she interfaces with large language models and her YouTube audience and the current Zeitgeist of the moment's news, she brings not just the sum of her formal education but the integrative insights, irony, humor and wisdom that come from three decades of cross-pollinating across these fields bringing more than the sum of parts to the synthesis. Similarly our neural net artificially intelligent trained models currently provides rapid and superhuman instant access to vast stores of information with steroid charged pattern recognition and analogical and logical capabilities; the human provides both experiential and embodied conceptual frameworks that make this information meaningful but also the gloss to alter, enhance and augment the information given and the ability to recognize and focus which patterns matter in a conversation which can also become Hegelian dialectic or thesis/antithesis synthesis between human and machine.
This synthesis becomes particularly evident in "resonant interactions"—moments when human insight and machine capability amplify each other. Consider a researcher using an AI system to explore potential drug compounds. The system can evaluate millions of molecular combinations, but it's the researcher's intuitive understanding of biochemistry—built through years of laboratory work—that helps guide the search in promising directions. The AI provides breadth, the human provides depth, and in their interaction, they achieve something neither could alone.
We might even find new terms for these such as 'cybergenetic resonance'— or interactions or moments between human insight and machine capability which open opportunities for emergent patterns of new discovery solution spaces. The AI system provides vast computational depth in analyzing and screening millions of molecular combinations, while the human expert contributes something more subtle yet equally crucial: the interpretive gloss and intuitive discriminating choices and navigation that illuminates promising directions within these complex solution spaces. Consider a researcher using an AI system to explore potential drug compounds. While the AI can map and analyze enormous molecular landscapes and territories screening towards promising directions, it's the researcher's experiential gloss—their intuitive understanding of biochemistry built through years of laboratory work, study and experience—that also helps identify which regions of this vast solution space merit deeper exploration. The co-intelligence here is a conversation and dialogue, what the social philosopher Bakhtin would call, the dialogic imagination, here this dialogue's efflorescence between human and machine producing discovery, invention and insight.
As these human/ai interactions begin to proliferate and repeat and then flower across multiple researchers and systems, they may create what we might term 'cybergenetic fields'—fertile solution spaces that also find larger synthetic formalization and compound through the accumulated wisdom of human-AI interaction and these possibilities for co-intelligence. The human's intuitive pathways of investigation help illuminate and navigate these interdisciplinary or multidisciplinary fields, discovering synthetic insights that might otherwise remain hidden in the deep computational space and disciplinary domains that have largely remained siloed because of the lack of dialogue even within what would be termed research institutions or collectives like the venerable US Ivy League and this associated disciplinary strictures and what has been allowed and not allowed. It is wise here to remember too that heterodox thinking like Einstein's or a host of other thinkers unmentioned here was not overly welcomed at first by Academia and many of Einstein's most venerable discoveries were made not from university ivory towers but rather as a lowly patent clerk as these positions were not granted or preferred for candidates like himself. With AI a new renaissance or possibility for discovery and invention will occur with current affordances. Like a seasoned explorer who knows which valleys might lead to hidden passages, a wider berth of human's experiential gloss can help chart new and amazing meaningful paths taking advantage of 'the human in the loop' and possibilities for AI's vast analytical possibilities and new territory opened.
The new AI/human interaction possible isn't simply additive—it's transformative on paradigm shifting levels. Our AI's now can construct deep computational spaces of complex hyperdimensional analogies and connections between disparate disciplines that up to now have been impossible to see or develop. Human expertise and intuition can for the moment provide excellent navigational wisdom individually or in large multidisciplinary and interdisciplinary teams to provide and confer very valuable interpretive gloss that will transforms raw data landscapes into meaningful discoveries and perhaps more Nobel prizes for all of our subsequent years into the 21st century. Together, humans and AI will create something richer than either could achieve without each other: solution spaces for insight, discovery, invention and solutions to challenges and problems that are computationally deep and intuitively navigable."
Yet this marriage of silicon and sapience remains in its infancy. The true potential of cointelligence and our human gloss and intuition lies not in current achievements but in future possibilities: systems that can learn not just from data but from the quality of human attention, which can incorporate not just explicit knowledge but the subtle patterns of human intuition. An important paper in the development AI's and large language models is titled 'Attention is All You Need'. The paper developed the idea of AI's Transformer architecture based on a concept of 'Attention' divided into Keywords (Keys), Queries (Questions) and Values - essentially what to pay attention towards - for now, this is determined by the human in the loop. As AI's develop multimodal capabilities, multisensory and multimedia, now envisioned through humanoid robots being embodied in the world and empowered by AI Imagine an AI system that could learn not just from Einstein's papers but from his way of seeing the world, that could internalize not just the facts of human experience but its felt texture, sounds, objects and spatio-temporal orchestration.
The metaphor of human/ai marriage, cyborg, centaur, co-intelligence all proves particularly apt here, suggesting as it does both unity and distinction, intimate connection without dissolution of difference. Like any successful marriage, cointelligence will require both parties to evolve while maintaining their essential nature. The machine must become more intuitive without losing its computational precision; the human must become more systematic without losing the creative chaos, intuition and gloss that makes us human.
This evolution points toward what we might call "emergent insights, knowledge, discovery, invention or wisdom"—insights that arise not from human or machine intelligence alone but from their dynamic interaction. Just as marriage at its best produces not just offspring but a new way of being in the world, cointelligence promises not just enhanced problem-solving but new modes of understanding, new ways of knowing that transcend the limitations of both human and artificial intelligence.
The road ahead remains uncertain. We must navigate complex questions about the nature of consciousness, the role of embodied experience, and the relationship between information and understanding. Yet in this uncertainty lies opportunity. As we develop more sophisticated forms of cointelligence, we may find that the true revolution lies not in artificial intelligence superseding human intelligence, but in the emergence of hybrid forms of knowing that preserve the best of both worlds—the machine's tireless processing with humanity's hard-won wisdom, silicon's speed with emergent co and new sentience for both humans and machines.
In this light, cointelligence appears not as a temporary phase but as the beginning of a new evolutionary chapter in the history of mind and our adaptation to not just survive but thrive in our world and universe. This narrative will now be one written jointly by human and machine, each contributing its unique form and gloss to our understanding of our ever-evolving world.
#HumanAIInteraction, #Cointelligence, #AI, #LLM, #AIResearch, #Discovery, #FutureofWork, #Innovation