Crossposted from my Substack. This is a rough introduction to a series of thoughts I had regarding our interface with AI -I'm hoping to broaden it more in the future.
As Artificial Intelligence becomes ever more pervasive in our everyday lives, there becomes an overwhelming need to address how humans and AI connect and interact with each other. This also includes nebulous structures that we deem of “intelligence” and “consciousness” in general, as we attempt to identify them in non-human entities without a strict definition of it within ourselves. One particular example is the application of the term Artificial General Intelligence (AGI) as an important benchmark of progress - commonly defined as the point at which an AI is as capable as a human. I argue that there is no clear definitive way to prove this, and the timelines and predictions will continue to shift until we can ascertain what human intelligence is and if AI can match or surpass it. Throughout its life cycle, AI has always been effective at creating new ways to question the status quo of what it means to be human - an extremely practical application of Science Fiction to humanity in general. We strive to create intelligence either in our own image or merge with it to create a connection that creates stable and harmonious relationships without being fully aware of our own capabilities or functions. Therefore, I wish to approach this categorization from a different angle - creating a separate vocabulary for the ways in which AI engages with humans; the main framework of this is what I call the hypermanifest. This introductory piece aims to explore and define this - a new conceptual framework for understanding AI’s evolving role in human connection and interaction.
Breaking it Down
I aim to break down this term into two parts - the concepts of Wilfrid Sellars’ Manifest image and Jean Baudrillardian’s Hyperreal. Sellars introduces the manifest image as humanity’s initial conceptual framework, an understanding through direct, experiential human perception. This is in contrast with the scientific image, extending beyond immediate human experience and seeks to explain the world in terms of scientific, often unseen explanations. Sellars explains that:
The 'manifest' image of man-in-the-world can be characterized in two ways, which are supplementary rather than alternative. It is, first, the framework in terms of which man came to be aware of himself as man-in-the-world. It is the framework in terms of which, to use an existentialist turn of phrase, man first encountered himself—which is, of course, when he came to be man.[1]
David Chalmers in Reality + differentiates the manifest from the scientific image by stating that:
In the manifest image, we’re free and conscious beings whose actions result from reasons and decisions. In the scientific image, we’re biological organisms whose actions result from complex neural processes in our brains [...] In principle, we can distinguish the manifest image of the Sun (the Sun as we think of it in ordinary life) with the scientific image of the Sun (the Sun as science reveals it to be).[2]
Frontier models such as LLMs (Large Language Models) process datasets that incorporate a blend of both manifest (like social practices, art and language) and scientific information. This allows LLMs to adopt characteristics of the manifest, developing their own interpretive models (some may even say internal world models). These models, through generating these responses, craft a unique reality in their dialogue - echoing the surface-level aspects of human interaction without accessing its deeper meanings. This is also akin to Baudrillard’s hyperreal; he introduces this concept as a state where the distinction between the real and simulated blurs, leading to a realm where representations can become more “real” than the things that they once represented. In this way a new reality is created, predominantly defined from the interplay of simulation and perception. In Baudrillard’s Simulacra and Simulation, he states that “abstraction is no longer that of the map, the double, the mirror or the concept. Simulation is no longer that of a territory, a referential being or a substance. It is the generation by models of a real without origin and reality - a hyperreal.” We can even see this quite literally as Large Language Models generate a response from an artificial perspective - not an inherently negative way, simply different. The choice of the word hyperreal is a deliberate one, in that Baudrillard claims that this transforms rather than replicates the original:
The real is produced from miniaturised units, from matrices, memory banks and command models—and with these it can be reproduced an indefinite number of times. It no longer needs to be rational, because it no longer measures itself against either an ideal or negative instance. It is no longer anything but operational. In fact, it is no longer really the real, because no imaginary envelops it anymore. It is a hyperreal, produced from a radiating synthesis of combinatory models in a hyperspace without atmosphere.[3]
If we analyse this quotation in particular, he purports that the real is already commodified into data (matrices, memory banks etc.), into smaller units that can be reproduced without measure or degradation. It does not need to be validated or compared to another existing form of reality as it is operational in its own right. It is a construct that exists outside of the boundaries or realms of the physical, as well as independently of the constraints that govern the once “real” world. In relation to AI, this can relate to the way that machines consume, learn and generate responses built on this data - creating a world or model of reality that is functional and real in terms of being operational, but is not grounded in experiential reality. This combines the structures of both the manifest (the input) and the hyperreal (the output). I propose that this combination is the foundation of the hypermanifest - imagined conceptually as the interface or latent space where human and AI connect.
The Implication of the Hypermanifest
If we look to how examples of AI such as predictive text, social media algorithms and GPS have already shaped the way we not only receive and share information but how we perceive space and time in terms of interaction, these examples would be the tip of the iceberg when it comes to the scale of the hypermanifest. As AI becomes more adept at automating tasks, gaining agency and understanding of human preferences, the emphasis of this new construct of reality is on functionality over fidelity. This concept of operative reality could have an effect in terms of healthcare, education and understanding of the world as we know it through scientific and technological breakthroughs. For reference, Vernor Vinge mentions in his well renowned essay The Coming Technological Singularity - How to survive in a Post-human era in 1993 that:
Another symptom of progress toward the Singularity: ideas themselves should spread even faster, and even the most radical will quickly become commonplace [...] when I began writing science fiction in the middle 60s, it seemed very easy to find ideas that took decades to percolate into the cultural consciousness; now the lead time seems more than eighteen months.[4]
The lead time today would be a fraction of this; information and discoveries fluttering in and out of our feeds and consciousness daily. As compute technology becomes more sophisticated, energy efficient and capable, coupled with an increasing abundance of opportunities to interact with AI, this stream of information and ways of interacting will continue to increase.
I envisage the hypermanifest as not only as a realm generated by AI but as an interface, a latent space in which humans and AI interact with each other, creating a new sense of reality through a new configuration of spatiotemporal conditions. I will illustrate this by using Doreen Massey’s model of relative space - defined by the actions and movements of agents both human and non-human without needing the containment of physical borders or time. She makes three propositions with which we can envisage this:
First, that we recognise space as the product of interrelations, from the immensity of the global to the intimately tiny [...] Second, that we understand space as the sphere of possibility of the existence of multiplicity in the sense of contemporaneous plurality as the sphere in which distinct trajectories co-exist [...] if space is indeed the product of interrelations then it must be predicated upon the existence of plurality [...] Third, that we could recognise space as a simultaneity of stories-so-far.[5]
I believe that the hypermanifest will become a mediation of our lived world as human and AI merge - either literally (as we’ve seen with the recent breakthroughs of Neuralink on the 20th March 2024 where their first human patient implanted with its brain-chip was shown controlling a mouse to play online chess using their thoughts) or inside this latent space where multiple trajectories coexist and shape its fabric. This shift contains a realm of possibilities that are both empowering and threatening, with risks that could occur in terms of personal privacy and autonomy. I hope to continue this series with the implications of the hypermanifest in more specific use-cases in the future. Hopefully this provides a different and interesting perspective in the ongoing discussion of human and artificial intelligence.
Bibliography
Baudrillard, Jean Simulacra and Simulation (US: University of Michigan Press, 1994 ed)
Chalmers, David J Reality +: Virtual Worlds and the Problems of Philosophy (UK: Penguin Random House, 2023)
Massey, Doreen For Space (London: Sage Publishing, 2005)
Sellars, Wilfrid Empiricism and the Philosophy of Mind (London: Routledge & Kagan Paul Ltd)
Crossposted from my Substack. This is a rough introduction to a series of thoughts I had regarding our interface with AI -I'm hoping to broaden it more in the future.
As Artificial Intelligence becomes ever more pervasive in our everyday lives, there becomes an overwhelming need to address how humans and AI connect and interact with each other. This also includes nebulous structures that we deem of “intelligence” and “consciousness” in general, as we attempt to identify them in non-human entities without a strict definition of it within ourselves. One particular example is the application of the term Artificial General Intelligence (AGI) as an important benchmark of progress - commonly defined as the point at which an AI is as capable as a human. I argue that there is no clear definitive way to prove this, and the timelines and predictions will continue to shift until we can ascertain what human intelligence is and if AI can match or surpass it. Throughout its life cycle, AI has always been effective at creating new ways to question the status quo of what it means to be human - an extremely practical application of Science Fiction to humanity in general. We strive to create intelligence either in our own image or merge with it to create a connection that creates stable and harmonious relationships without being fully aware of our own capabilities or functions. Therefore, I wish to approach this categorization from a different angle - creating a separate vocabulary for the ways in which AI engages with humans; the main framework of this is what I call the hypermanifest. This introductory piece aims to explore and define this - a new conceptual framework for understanding AI’s evolving role in human connection and interaction.
Breaking it Down
I aim to break down this term into two parts - the concepts of Wilfrid Sellars’ Manifest image and Jean Baudrillardian’s Hyperreal. Sellars introduces the manifest image as humanity’s initial conceptual framework, an understanding through direct, experiential human perception. This is in contrast with the scientific image, extending beyond immediate human experience and seeks to explain the world in terms of scientific, often unseen explanations. Sellars explains that:
David Chalmers in Reality + differentiates the manifest from the scientific image by stating that:
Frontier models such as LLMs (Large Language Models) process datasets that incorporate a blend of both manifest (like social practices, art and language) and scientific information. This allows LLMs to adopt characteristics of the manifest, developing their own interpretive models (some may even say internal world models). These models, through generating these responses, craft a unique reality in their dialogue - echoing the surface-level aspects of human interaction without accessing its deeper meanings. This is also akin to Baudrillard’s hyperreal; he introduces this concept as a state where the distinction between the real and simulated blurs, leading to a realm where representations can become more “real” than the things that they once represented. In this way a new reality is created, predominantly defined from the interplay of simulation and perception. In Baudrillard’s Simulacra and Simulation, he states that “abstraction is no longer that of the map, the double, the mirror or the concept. Simulation is no longer that of a territory, a referential being or a substance. It is the generation by models of a real without origin and reality - a hyperreal.” We can even see this quite literally as Large Language Models generate a response from an artificial perspective - not an inherently negative way, simply different. The choice of the word hyperreal is a deliberate one, in that Baudrillard claims that this transforms rather than replicates the original:
If we analyse this quotation in particular, he purports that the real is already commodified into data (matrices, memory banks etc.), into smaller units that can be reproduced without measure or degradation. It does not need to be validated or compared to another existing form of reality as it is operational in its own right. It is a construct that exists outside of the boundaries or realms of the physical, as well as independently of the constraints that govern the once “real” world. In relation to AI, this can relate to the way that machines consume, learn and generate responses built on this data - creating a world or model of reality that is functional and real in terms of being operational, but is not grounded in experiential reality. This combines the structures of both the manifest (the input) and the hyperreal (the output). I propose that this combination is the foundation of the hypermanifest - imagined conceptually as the interface or latent space where human and AI connect.
The Implication of the Hypermanifest
If we look to how examples of AI such as predictive text, social media algorithms and GPS have already shaped the way we not only receive and share information but how we perceive space and time in terms of interaction, these examples would be the tip of the iceberg when it comes to the scale of the hypermanifest. As AI becomes more adept at automating tasks, gaining agency and understanding of human preferences, the emphasis of this new construct of reality is on functionality over fidelity. This concept of operative reality could have an effect in terms of healthcare, education and understanding of the world as we know it through scientific and technological breakthroughs. For reference, Vernor Vinge mentions in his well renowned essay The Coming Technological Singularity - How to survive in a Post-human era in 1993 that:
The lead time today would be a fraction of this; information and discoveries fluttering in and out of our feeds and consciousness daily. As compute technology becomes more sophisticated, energy efficient and capable, coupled with an increasing abundance of opportunities to interact with AI, this stream of information and ways of interacting will continue to increase.
I envisage the hypermanifest as not only as a realm generated by AI but as an interface, a latent space in which humans and AI interact with each other, creating a new sense of reality through a new configuration of spatiotemporal conditions. I will illustrate this by using Doreen Massey’s model of relative space - defined by the actions and movements of agents both human and non-human without needing the containment of physical borders or time. She makes three propositions with which we can envisage this:
I believe that the hypermanifest will become a mediation of our lived world as human and AI merge - either literally (as we’ve seen with the recent breakthroughs of Neuralink on the 20th March 2024 where their first human patient implanted with its brain-chip was shown controlling a mouse to play online chess using their thoughts) or inside this latent space where multiple trajectories coexist and shape its fabric. This shift contains a realm of possibilities that are both empowering and threatening, with risks that could occur in terms of personal privacy and autonomy. I hope to continue this series with the implications of the hypermanifest in more specific use-cases in the future. Hopefully this provides a different and interesting perspective in the ongoing discussion of human and artificial intelligence.
Bibliography
Baudrillard, Jean Simulacra and Simulation (US: University of Michigan Press, 1994 ed)
Chalmers, David J Reality +: Virtual Worlds and the Problems of Philosophy (UK: Penguin Random House, 2023)
Massey, Doreen For Space (London: Sage Publishing, 2005)
Sellars, Wilfrid Empiricism and the Philosophy of Mind (London: Routledge & Kagan Paul Ltd)
Vinge, Vernor The Coming Technological Singularity - How to survive in a Post-human era https://ntrs.nasa.gov/citations/19940022856 <Accessed Oct 30, 2023>
Wilfrid Sellars, Empiricism and the Philosophy of Mind (London: Routledge & Kagan Paul Ltd), p.6
David J Chalmers, Reality +: Virtual Worlds and the Problems of Philosophy (UK: Penguin Random House, 2023), p.425
Jean Baudrillard, Simulacra and Simulation (US: University of Michigan Press, 1994 ed) p.1
Ibid, p.2
Vernor Vinge,The Coming Technological Singularity - How to survive in a Post-human era https://ntrs.nasa.gov/citations/19940022856 <Accessed Oct 30, 2023>