! Thank you very much. Thank you, Jean, for inviting me. Thank you, Christina. It's been a wonderful loss. Thank you for the students who came and stayed with me for three days. We had a very great discussion every day. I have prepared something for you to follow me, if you can't follow me otherwise. I know things are challenging and this is a challenging topic, just to think with machines or about machines is challenging itself, although we use machines all the time.
So my talk is really to try and rethink of cybernetics, which is a science of communication and information which was born during the 1940s. It suggests that since the 1940s, a philosophy or the technological project of Western metaphysics has been reconfigured by information technologies. Our age, as Giorgio Agamben, an Italian philosopher, argues, has radicalised what Guy Debord called the society of the spectacle, because the mediatic regime of contemporary capitalism has tapped into the very communicativity and linguistic being of humans. According to Agamben, what human beings communicate to one another is pure communicability,
that is, language, and politics precisely arises as the communicative emptiness in which the human emerges as such. In particular, in this age of communicability, we shall be concerned with what is the matter of thought itself, or what is the power of thought. This power, for Agamben, means to be into language itself, as pure mediality, being into a mean, as irreducible condition of human beings. But this immersion into mediality also reveals that with cybernetics, the very automation of communicability, the automation of language,
also thought and also thought has been really challenging what for Western metaphysics is philosophy. From the automated networks of decision making in governance and security, to the unsupervised routine of intelligent logistics and the technological reinvention of gender, race and class and ultimately of the category of the human, cybernetics has embodied the mediality of communication, the mediality of language in terms of information, information retrieval, information storage, information transmission. Here cybernetics shall be not simply seen as a science of communication and control in prediction or governance of population,
but as part of a more general historical reconfiguration of metaphysics on behalf of techniques, i.e. on behalf of machines, where instrumentality has led to a profound transformation of what we take the social to be, and also what we take the very effort of thinking. In short, I want to suggest that with cybernetics, technicity, instrumentality, machines, has finally shifted away from a role of just demonstrating an idea and has acquired a quality in which it can articulate in a new way the relation between means, i.e. instruments, and ends, i.e. ideas for which instruments need to work.
cybernetics as forced I would argue Western metaphysics to abandon is teleological project of reasoning and ultimately has brought thought to to confront the consequences of thinking through a weird machines Martin Heidegger in 1969 wrote this little essay called the end of philosophy and the task of thinking. For Heidegger, philosophy and the teleology of Western metaphysics has evolved into scientific knowledge and technology. A science and technology during the Second World War took over and replaced humanities and philosophy for him. And this kind of mentality
has to do with the effectiveness of results and the ultimate replacement of theoretical reasoning with the supposition of categories. But I want to question Heidegger's critique of cybernetics, which for him demarcates the point of completion or the ultimate realization of a Western technological thinking when communication science replaced the so-called rule of reason with instrumentality or thinking. Since World War II, cybernetics and computation have imparted a new image of reasoning, related not just to the mechanical behaviour of clocks, but to the responsive feedback and the prediction activities of intelligent machines.
We are all used to the responsive feedback of our little computational machines that we use today, namely the mobile phone and social media. With cybernetics, reasoning was replaced by a dynamic form of calculation, embedding humans and machines in endless feedback loops bound to information signals, noise, recording and transmission of data. Here learning became central to the task of thinking, which replaced truth, truth, i.e. reasoning based on judgment by discernment with a procedural process of understanding by means of testing, trial and error.
The scientific reformatting of philosophy did not grant reason a universal schema of rules from which one could deduce truth. Instead, the alignment project of scientific emancipation had unintentionally entrapped its very principle, i.e. reason, within the science of organisation and governance of labour, the arts and all aspects of social life. With cybernetics, Western metaphysics had renounced the security of concept deducing truths and entered an automated world of formula, regulatory cause and probability calculation. In other words, as science became the instrumental
function of reason, they would purge philosophy from myth. Cybernetics also stripped metaphysics from some kind of theoretical idealism, showing that rules-bounded thinking follow instead an inductive mode of trial and error, a way of discovering unknowns, rather than confining thinking to what is already known. With cybernetics reasoning becomes instantiated in automated learning, self-regulation of behavior that adapts to the environment and modify its initial premises. So for Heidegger, cybernetics announced the end of Western Metaphysics, the reduction of metaphysics
and philosophy to instrumentality. It reconfigured the nature of governance through some kind of efficient machines based on algorithmic communication. But soon after the 1940s and during the Cold War, the cybernetic infrastructure of governance was still functioning as some kind of technological unconscious of modernity. because what technology was doing was actually acting as a symptom of failure, of the failure of the modern project of rationalisation, granting some kind of an unlimited exploitation of the world with the Industrial Revolution. Techniques then were still embodying this kind of unconscious of philosophy,
despite eliciting a new aesthetic and a new political vision of a dehumanised culture. By acting as the mediatic unconscious of this kind of informational infrastructure of thinking since World War II, cybernetics therefore was the system for governance or has become since then the system of governance of the social, relying on truth carried out by machines. carried out by machines so for I the girl deficiency of information was also the determination of man is enacting social being for it is the theory of the steering of the possible planning and arrangement of human labor
cybernetics for him transforms language in an exchange of news similarly the arts become regulated and regulating instruments of information. With the migration of cybernetics into the area of computation and artificial intelligence, logos or reason became enmeshed with the ratio or calculus. As historians of science Lorraine Dawson explains, after World War II, the replacement of the faculty of judgment with algorithmic rules rules turn reason into rationality, the fitting of means to ends, based on statistical prediction and correction of human irrational behavior. But this
argument was already confuted in the 1980s, when post-critical understanding of cybernetics argued that the techno-rational governance had already triggered the crisis of Western reasoning, because the mediality of thought had finally exposed the political dimension or the political tension between theory and practice, and thus challenged the colonial, the patriarchal project that imparted its truth on the instruments, the bodies, the population, the races and genders that it used to obtain its pre-ordered results. By exposing the tension between means and ends, patterns and randomness, rules and contingencies, truths and uncertainties,
what cybernetics did, therefore, is that it became the opportunity to develop a new critical thinking of the human subject. A critique of its organic wholeness and its cognitive representation of the world, therefore challenging the representational perspective of the material, the affective and the bodily consistency of the real. For instance, Donna Haraway in the 1980s wrote the Cyber Manifesto in which she argues that the cyborg defined and challenged the modern rational subject whose mission against myth, superstition and obscuratism had led to the violence of colonialism,
patriarchy and capitalism. But the consequences of this modern rational subject project was that its techno-progressive vision instead had led to a realisation that that the technicity had come to coincide with the non-pure and contaminated cultures claiming for the intersectionality of gender, race and class and also of animal, human and machine. Importantly, the domination of rational instrumentality was here reversed. The instrument did not follow Asimov's law and instead of simply rebelling against the The master, it pushed to its extreme consequences the already debilitated master rule of reason,
as this entered the recursive patterns of information. Here, technicity can no longer remain unconscious and invisible. Finally, technicity enters the realm of a no longer human consciousness. In other words, techniques had itself become instrumental to consciousness, but not simply because, as Adegaard claimed, it came to replace Western metaphysics. Instead, what I would like to suggest, what I've been trying to work on, is the fact that the idea that if instrumentality has become central to consciousness, it is because it offers a possibility of working through the limits of the metaphysical project of reason,
that is, of exposing the consequences of instrumentality that corresponded to the social political claim for the origination of another mode of reasoning, moving from within and throughout the meaning of instrumentality. This vision of instrumentality, however, is not simply an invitation to approach technicity or machines as a way of crafting the world away from logical machines or to argue for some kind of nature that we have to go back to and that could reformat or help us to rethink this domination of information on the planet. Instead, my attempt is to rethink critically and constructively instrumentality as the means by which Western metaphysics had to work through what he had kept most invisible from itself,
i.e. the limits of thinking insofar as the very instrument upon which the transmission of truth was based rather revealed a general condition of indeterminacy of fallibility irrational thinking in short this is an effort to theorize the artificial thinking of machine as the origination of an alien thinking of an alien reasoning which does not just correspond to the debilitated subject of Western patriarchal a wise a wise subjectivity instead what remains crucial to unpack is the possibility of alternative modes of thinking think triggered by this emergence of instrumentality where in uncertainty
place instead technique beyond the dominant image of cybernetic control. So my attempt here is to reverse the critical technology and rethink technology as the true cybernetics as a way to bring forward another mode of reasoning, a mode of reason that was born through historically during the 40s and after during the Cold War as a mode of challenging the modern project of rationality through which colonialism, patriarchy, and capitalism had activated the teleological project. So I am going to refer to two models of critique of cybernetics and computation
that have been kind of arguing about how can we rethink technology away from power or how do we think of technology, this mode of computational infrastructure that we have today in terms of power. So for instance, I'm going to look at what Brian Massoumi calls onto power and Benjamin Bratton calls computational sovereignty because both these models both of these critiques seem to ground the end of reason into some kind of metaphysical machine machinism without soul a kind of understand the technology where the instrument looks back at us indifferently instead mind is some kind
of an attempt of some kind what we could call a philo fictional attempt to argue for the historical origination of an automated reasoning in terms of a denaturalized metaphysics thus involving a logic and the knowing stemming from the cybernetic inquiry into thinking and also the cybernetic inquiry into becoming the instrument of thought I turn to some a pragmatist understanding of instrumentality especially in particular John Dewey to suggest that information is not simply a means of communication. Instead, the instrumental or practical use of information is this?
Sorry. Practical use of information also involves an inferential mode of reasoning, involving the generation of meaning and not just the demonstration of truth. Insofar as this automated system that we are dealing today, especially modes of thinking such as machine thinking, I don't know why it's not moving, actually are not just matching data, but they are learning from data. Whilst autonomous learning has always been central to both cybernetics and computational envisioning of intelligent behavior, post 1980s expert and knowledge system, a knowledge system, intelligent agents, machine and deep learning intelligence automata,
shall rather be considered as part of a third order cybernetics. So my question is whether we can offer an alternative articulation of artificial reasoning and thinking of the instrument and instrumentality not just as a mean to an end, but as something that can engender itself a logic. The larger implication of this metaphysics requires a further analysis about the political implication for the post-1980s radical claim about the ontological alliance between gender, race and machine. In short, my question is, what kind of instrumental reasoning can be used against and away from
the totalizing image of cybernetics as the triumph of mediality or the collapse of logical thinking, reasoning and metaphysics? So, Massoumi's analysis of neoliberal and neoconservative governments discusses control in terms of what is called second order cybernetics. An echoing Michel Foucault argues for the becoming environmental of cybernetic power. What does it mean? That governance today coincides with a continuous regulation of effects, the automation of responsive behavior acting as checkpoints of verification beyond the appeal to pre-established causes. At the core of this new form of automation is the tension between information patterns, noise and randomness,
which following Basumi define what he calls on top power I the pre-empting of uncertainty power is haunted by unknowns or ontological liminal thresholds between its interactive functions and the visual field of its action the modus operandi so cybernetic decentralized interactive and parallel procedures overlaps with the temporal and affective field of variations of conduct. For Masumi, cybernetic control shows that the rational system of prediction has become attuned to the visceral recording of continuous effects, replacing rationality with affectivity.
Masumi calls this activity of prediction pre-emptive strike. Environmental governance indeed replaces statistical classification by directly monitoring effective responses and learning from data in order to measure unknowns. Cybernetic control here espouses the metaphysics of indeterminate causality coinciding with the snowballing effects with a process or a behavior of nature here causality coincides with the condition of the accident or indeterminate indeterminacy this is why he says the power today works as an environmental form of preemption not a rational logic aiming at in framing the
truth on the future but some kind of effective piercing of the future through the cybernetic vision of the present. Instead of a cognitive rational mapping, cybernetics control depends upon the affective responsive to potential threats and thus subsume cognitive reasoning not only to affect the procedure but also to a zone of indistinction between probability and potentiality. From this standpoint power is not an ideology but is continuously produced and presupposed induced and deduced active and passive and operates in this zone of logical indistinction between being and control but this view of cybernetic control seems
to preclude a discussion about a less clear tendency in computation i eat the the fact that with the shift from reason to procedural reasoning that already happened during the Second World War and the Cold War, what also this transformation allowed was not just a kind of denigration of reasoning, a one-in-down of reasoning, but actually the emergence of a new mode of thinking through procedures. This means that, for instance, if we go and look at what Benjamin Bratton calls the planetary computation, it shows instead that what governance operates today, the way it operates today,
is through a cybernetic megastructure that has engendered its own perceptual and conceptual mode of visions. It's not just us who look at machine, it's also machine that look at us. So this kind of historical transformation of cybernetics during the second world war and then cold war during the 80s has led to a kind of automation of thinking by through machine. But this thinking is not just about procedures or tasks that are added one against the other to give us results. It's also the capacity of machine to think and to preempt or predict uncertainty or unknowns or to impose a certain mode of conduct or some kind of envision or inscribe a certain mode of conduct.
And this is what, for instance, Benjamin Bratton calls in order of, in his kind of visualization of how this kind of infrastructure of this automated system work. you know, if you look at logistics, branding, of course the internet, security, social networks, there is some kind of intelligence operating within and beneath the surface. So there is this kind of disconnection between what we think we are operating and what we think the way we are using machine to achieve a task And instead the way algorithmic intelligence is using data to achieve other tasks that are not necessarily transparent to us.
Actually it's the opposite. So in order to describe this kind of governance, cybernetic governance, Bratton uses this form or this concept of the stuck in order to describe the new geopolitical orders which is both indifferent to the aesthetic and logical sensibility of the human. The stuck imparted some metaphysical principles of social governance that are mainly subjected to the internal contingency of this mega-structural cybernetic control. Because obviously what has become important is that from the 18th onwards,
is that actually the idea of machine efficiency has been completely confuted by the idea that actually machines break down. There is an indeterminacy in the way logic in machine operates. So this means that there is contingency. The machines are open to be affected by unknown causes. So the idea that the Second World War idea the machine could actually replace reason or the kind of human error with some kind of mechanism that could purge function from any kind of error, error from any kind of mistake has completely been turned over after the 80s,
where instead there's been a recognition that the limit of computational logic and the limit of cybernetic functioning is exactly fallibility, lack of completeness of data, and what is also being called the incomputable. So there is this kind of interesting transformation that help us also to think that if we were to say, well, machine can, the critique to technology needs to be based on some kind of recuperation of social, cultural, you know, social cultural critique that, whereby machine could never be able, are just some kind of cold rational system, this kind of critique doesn't hold anymore
because machines are not just cold rational system because machine, the kind of computer, developing in artificial intelligence and in cybernetics has instead led us to think the machine can also think the unthinkable, i.e. that they are able to work through breakdown, through fallibility and through contingency, which is something that was not what was believed when they were actually invented during the 50s, 60s and 70s, so when they were actually, there was this other mentality of the efficiency of the machine. So what Benjamin Bratton is actually also pointing out is that there are contingencies within this mega-structural machine that we are embedded in today.
He talks about the stack as the state, and he says that the stack is a machine, and the machine is the state. And it's not just representative of power, but it's generative of power. is generative of perception, cognition and action that are mediated by computational logic, by cybernetic feedback of communication and interaction. Accidents and not logical reasoning are the ontological condition of this form of governance. Local computations are here contingently reassembled in an evolving mega-machine whose sources are human discourses, human bodies, affectivities, emotions and behaviour.
And basically what he's saying is that as we deliver our data to data systems whenever we are purchasing a book or we are buying a ticket to go to the theatre or we are passing border controls, what we are doing is provide data to this evolving mega machine whose sources exactly are human discourses, the way we talk, we communicate through machines, human bodies, the data about your health for instance, affectivities, the way we empathize or not with other people, emotions and behavior, as well as information on molecular, chemical, physical, cultural, social and personal skills.
So if second order cybernetic was concerned with this kind of system of affective control, control, the stack is closer to a third order cybernetics because its metaphysics is not without a cause. No longer some kind of rippling effect that stops the machine from total control, but the conscious realization or the interiorization of the contingency of mediality, i.e. to communicate even in even in the form of automation is a contingent affair. You always are dealing with the possibility of error and also you're dealing with unknown data.
So the idea that this mega machine is in control, is a totalizing mega machine, does not correspond to reality because the mega machine, on the contrary, function precisely because it's able to deal with uncertainty, non-predictable behavior and unknown. So in a way indeterminacy or uncertainty is at the base of this machine. So when we think about automation we are used to think about the efficiency of a mean that gives us, that carries us, allows us to carry out a task. For instance, here, a great example, the HAL of 2001 Space Odyssey is a computational machine
that is not used to error. And when it sees error, i.e. human, that is the human, it stops working and or the task is to eliminate the error. The kind of machine we're talking about today, i.e. machine learning, deep learning, Google Dream, what all the machines that are post post 80s and post 90s are really machines that instead do work through error so error or is no longer a limit to computational and cybernetic thinking and that's also what Benjamin Bratton is trying to describe and argue that this is therefore a new mode of dealing with governance and with power that has to account for this kind of fragility
vulnerability accident determined decision-making machine so like the computational looping branching a processing of the algorithm values entered into an interspace the stack entails sovereign decisions based not on fixed protocols but on the accidental interaction between layers accumulating into more comprehensive structure. The stack is already crafting politics, geography and territory into the image of an accidental decision machine. I think this is also something that we can grasp quite intuitively if we think about whenever we want to try to get something done and we are the reply that we get is the computer says no. So there's this kind of, if there's an accident, the
computer is down, the system is down. This is also something that has justified a lot of amounts of money that has disappeared or reappears all of a sudden when, for instance, there are transactions with high frequency trading. there is this kind of vulnerability and contingency of the machine that allows precisely for a new governance to operate, to operate by error rather than to operate by excluding error and that's an interesting shift that Benjamin talks about in terms of how not only automation of labor but it's precisely the
automation of perception and cognition that is able to articulate this kind of multi-layer infrastructure of software protocol where network technologies operate within a modular and interdependent order exposing the internal contingencies of cybernetic metaphysics fragmenting this totalizing image of procedural efficacy into thousands of computational machine or computational procedures at all scales. It's also the idea that there is not one centralized computational power. It's precisely based on many many layers of autonomous computational processing of information of all the data that can be
gathered by by system of governance or institutions or health services or insurance or banks which most of the time we deliver in a very spontaneous way to put it crudely whilst the idea of the onto power the Masumi describes explains cybernetic governance as some kind of continuous modulation of affects preempting the future, stopping the future from happening, the stack instead envisioned the fact that cybernetic machine thinking is made of contingency and that's where complexity lies. These views thus offer us a very important contribution, at least for my research and hope I can communicate
that to you, to reframe instrumental reasoning away from this kind of Adegarian conviction that cybernetic fulfils the dream of Western metaphysics. But my attempt is rather to re-envision cybernetic metaphysics, not just as some kind of accidental or some kind of contingency-led system. Because what I'm interested in is to recuperate a mode of reasoning or to enlarge the category of reasoning to non-human modes of processing information. So mine is a kind of critique of human sapiens but also rethinking of what human sapiens
or human reasoning, how to rethink it by starting with looking at how the means, how the means, the instrument, has indeed developed a mode of reasoning through cybernetics and computation because what has happened with cybernetics and computation is that there has been this shift from reason as some kind of truth to reasoning as learning. So there is another mode of rethinking not only human reasoning, but also human subjectivity that therefore human politics has been open to the instrument that has been used by Western
metaphysics and has rebelled against Western metaphysics and has produced a new form of reasoning which seems to me quite urgent, we engage with and try to unpack rather than then assume there actually is a problem of science or is a problem of governance. This is also a problem from art, for humanities and for politics. So my attempt is to situate the origination of instrumental logic within the scientific investigation of the relation between information patterns and randomness.
Here algorithmic functions learn to transcend themselves not in an accidental fashion, so not beyond reason, because to admit that this mega machine operates beyond reason is to admit a political defeat as to how instead the task of a critical thinking of machine or technology should be instead rather to engage or to unpack or to formulate modes of logic that are inclusive of what we don't understand rather than being rather than being kind of arguing for a limit of reasoning
in machines. So, whilst this mode remains an alien or denaturalized form of reasoning and is neither completely idealistic nor empirical nonetheless it shall not be unaccessible to the arts and the humanities ontological and epistemological inquiries into the active thinking of machines. Instead, if the structure of governance continues to grow in this seamless accidental fashion perhaps critical efforts shall be able to offer accounts of model reasoning that are alien to us, so as to address assumption about what instrumentality
and what this kind of mediatic infrastructure of sociality in which we are all embedded today is and can do. One way to do so may involve a critical approach to algorithmic information theory, because for me algorithmic information theory is what explains how cybernetics has stopped being just a mode of statistical calculation, but has become also a mode of learning and thinking, the incalculable. So since the 1940s, there's been experiment in cybernetics and computation that have concerned, not just, that have concerned the development of an inferential reasoning with machines,
involving learning or the use of unknown information. One is to go back to Alan Turing's discovery of what he called the incomputable, and Godel claims about incompleteness in formal thinking to suggest that the automation of reason was a way to experiment with the medium of thought and extend logic and statistics beyond universal truth and probability. Only in the 1980s that post-Turing and post-second order cybernetic focused on this relation between information and randomness, to experiment with the function of autonomous learning. So an autonomous elaboration of information beyond premises, which is now central to deep learning artificial intelligence
and their neural net infrastructure. This is a predictive, a computational evolving machine. And what is important is that after the 80s is that machine or the automation of reasoning or the automation of learning really became emerged with what is called the kind of introduction of time into thinking. That's why learning becomes something that is important because it's about how you move from cause to effect from premises to result. And what happens in the middle is an elaboration of data. But the elaboration of data can only occur in time or through time.
In fact, one of the most important invention or kind of development of what already for Newman elaborated in the 1950s is cellular automata. Cellular automata are just algorithm, sets of procedure that are programmed to evolve in time. But this evolution in time is not just a linear evolution. This kind of cellular automata can evolve in a circular, non-linear, retroactive time. So there's always a way that they can map many modes of behavior or many modes of kind of patterning of data simultaneously at a long distance
and parallelly and so on. So time is what enters artificial intelligence. Not just some kind of representation of the mind through symbols, but kind of the very capacity capacity of change from premises to result, from input to output. And what becomes important for information theory and computation is that prediction, therefore, is no longer done on probability, i.e. on what was already known, or what you input into the system. Actually, it's the opposite. You start from somewhere to get somewhere else completely that you had not envisioned, i.e. the result cannot be contained in the premise, but needs to be, but the data contained in the premise
need to elaborate in time according to certain variables so to arrive to a result that you had not anticipated. And obviously this is, in art, it's been, you know, interestingly used to experiment with sound production that lasts a thousand years. for instance, you know, where the sound evolves and there is this generative algorithm that continues to evolve in time. Or, you know, in architecture where you have intelligent system or intelligent floor, intelligent walls that again are able to adapt to the input from the outside. But they are not just pre-programmed to do something specific.
they can also deal with what in information theory is called randomness. What is randomness? Randomness is not something arbitrary. That's something that can be chosen between two sets of information. But randomness is a volume of information that cannot be compressed into a single formula, into a number, or into some kind of, into a theory. So it's precisely the idea that incompleteness, i.e. randomness, is at the core of information. So when, for instance, when we talk about big data or data mining processes,
what we are talking about is machines that are trying to reduce, read or make a pattern out of data, because the quantity of data is infinite. Insofar as also the quantity of any quantities, the quantity of numbers is infinite, the quantity of realities is infinite. So there is this kind of realization in the 80s already with Gregory Chaitin, who is an algorithmic information theorist, who said actually what Turing discovered in the 40s, the idea of being computable, like you don't know when a program will stop, or you don't know whether results will ever match premises, you cannot completely have a formula
that can calculate everything, and he called the incomputable. He says actually this is at the very core of computational system today, i.e. computational processes are never ended, computational processes are never ended. And what does this mean? It means that indeterminacy or unknown or incomputable or massive volume of data are precisely the condition for calculation rather than being the opposite of calculation. So without infinite masses of data, you cannot calculate, you cannot predict. And even if you predict, even prediction can always only approximate the result,
but you can never know for sure. And this is why we have a lot of, you know, this is why we have this kind of fallibility into the machine. And you have this kind of things that are neural networks that are, you know, able to map different kinds of sites where information becomes more relevant or where data can be tracked or can be organised so that you have some kind of result, but nonetheless there is no complete computational system. But that doesn't mean, to my mind, that this computation system is therefore fallible breakable or in logic instead what my
question is that we rather look into or extend our notion of logic and reasoning and and and find out or kind of re theorize or challenge or if this is a big challenge, a challenge, our assumption of what logic, ideas, concepts and meanings are and rethink what or analyze and investigate or inquire into this logic of the instrument. Because this kind of instrument have a logic rather than being just something that operate magically or they break down accidentally
and because they operate through uncertainty, therefore they are themselves uncertain or irrational. I don't think this is the case and this is something that critically must be addressed. So from this standpoint, if cybernetic does not just mark the completion of Western metaphysics but the origination of an alien metaphysics or a technologic carried out by machines, it is because the historical transformation of the rule of reason into this kind of procedural reasoning involves the ingression of incomputables or uncertainties into modes of prediction and programming conduct. Incomputables or randomness
does demarcate the transcendental tendency within the medium to think beyond what it does. So what it does does not correspond to what it thinks because the system is so completely, this is a kind of surgical, the way that the use of machine learning in surgery, in brain tumor surgery actually. So here instrumental reasoning becomes also a mode of knowing how, it's developing another mode of knowing, enacting truth, enacting a certain kind of behavior and a certain kind of philosophy, a mode of knowing that is not contained within the idea
of human sapiens or the idea of rational thinking. So in order to address the computational strata of controlled societies today, it may no longer be sufficient to rely on this kind of given idea of reasoning or even computation. effort of inventing propositions about what the logical machine can become within our computational infrastructure requires therefore a rehabilitation of instrumental reasoning. My point is that if artificial thinking is acted as a denaturalized consequence of reasoning, i.e. de-essentializing universal truth, it can also become a space for inventing human politics, for inventing human
subjectivity through the alien reasoning of machines. One possible way to question the exceptionalism of human sapience without falling back into an anti-logical metaphysics could be precisely to start from what a John Dewey, pragmatist John Dewey called experimental logic. For John Dewey instrumentality should be concerned with the consequential elaboration of thinking from doing. So this is also a plea to say that when we look at machine learning or when we look at automated system or this intelligent automated system, what they are thinking does not correspond
to what they are doing. So it's not that they are unconscious or non-cognitive or they are they are just carrying out tasks. The question instead is that instrumentality itself, or the means, the action, the function, the carrying out task of a mean, already implies an elaboration of thinking from doing. Instrumentality therefore is not just what realizes an idea, but is a practical elaboration of truth. For Dewey thinking is itself an intelligent medium and similarly machines are intelligent techniques. Instrumentality therefore poses no ontological distinction between humans and machines. And yet one has to
argue that this commonality does not imply that the human and machine are of the same kind and the way their reason is equivalent. As activities, these particular practices of thinking do not have the same consequences or the same use of unknowns for achieving their logical reasoning, truth or axioms. This means that logic is entrenched with particular possibilities of working out what is known and what can be known, whereby particular in computabilities, particular modes of randomness are used not just as a limit to a system or a limit to logic but as an experimental eventuation of new premises and new
results. From this standpoint information is not simply the instrument of cybernetic metaphysics but is now used by learning intelligence as means, instrumentalities for machine knowledge. Far from defining a purposive drive of function after function, this pragmatist view of instrumentality accounts for an experimental use of unknowns as involving an elaboration of ends in terms of instrumental logic. But to argue for a philosophy of machines, because in effect that is what I'm trying to argue here, Whether we can not just say that automation, like Heidegger said, is the end of philosophy,
but actually with cybernetic what we have is a kind of transformation of reasoning through a bi-machine, through a bi-instrument. And this is for me also a kind of political twist twist that I want to give to the critique of the 1980s critique of cybernetics especially started with the work of Donna Haraway for which the cyborg was a figuration of a new mode of subjectivity whereby the natural idea of or the naturalized idea of gender, race, and class could be challenged through the instrument,
through information and communication technology. And for me this is just an elongation or an expansion of that project by also trying to figure out what kind of beast we are dealing with. Because obviously when we talk about, Donna Hara was talking about automated information system, communication system of the 80s, they were very different from the kind of post 90s technologies of communication that we have today that operate exactly by and through randomness, so by and through indeterminacy. and includes therefore modes of knowing that are of a different kind,
that are much more complex and much more able to elaborate their own thinking and not just something that can be reduced to a function. So, but to argue for a philosophy of machine is mainly for me a speculative effort that requires an attentive discrimination amongst practices of reasoning concerning not only how they work, but also how they conceptually can use randomness and generate new possibilities of meanings. What kind of meaning, what kind of knowledge, what kind of reasoning are machines producing?
This is also an effort towards the invention of methods where practical reasoning becomes central to work out and to work throughout and with general models that must revisit the separation between the philosophical project of the arts and the humanities and the cybernetic or technoscientific project of instrumentality. generality so my argument is for a general mode of knowledge that includes these efforts rather than separating them but also what it requires is a dialogical construction between premises and
results the invention of hypotheses to challenge assumption and to reassess our material practices can be collectively endorsed through the discourse of imagination and the multiplication of reasoning. So the effort that I am actually arguing for is an effort to rethink our technicity or instrumentality as a project that already started since the Second World War in terms of denaturalizing reasoning. because obviously the problem of a naturalized understanding of reasoning will just assume that there is a hierarchy within human culture
whereby humans have a faculty of thinking and reasoning as opposed to, and we know the hierarchies that were produced within the industrial period, these are the questions of race, questions of gender and questions of class. This kind of extension of the project of kind of mapping the evolution of instrumental thinking through time from the 40s when the argument is the reason had become mindless because it had already been substituted by information system and by tasks, as Heidegger had argued.
And then moving on to the late 40s when both the Turing machine and cybernetics emerged as a mode of wanting to mechanize reasoning. But then mechanization of reasoning showed that in reality, reason itself is constructed as a limit, i.e. is based on failure, that there is no possibility of having a system of reasoning that could be totalizing, it could actually include all possible results within its premises. So this was already unknowledged by Turing in the late 30s and the 40s and then activated in computational machine in the 50s.
My point is that what happens when this machine do not only instantiate reasoning, but start to think because they become experimental a vehicle for learning. And when learning becomes not only a way of repeating, reiterating something, but actually something that is already programmed in the system, but actually being able to use data to learn something else, to variate in learning, then you have another form of artificial intelligence. You no longer have all, but you have ex machina, for instance, in terms of cultural tropes, you have a kind of
machine that can communicate with other machine, that can develop cognition, affection, and also can, can also can develop a different mode of, asks us, or demands of us to rethink, to open up the idea of what does logic? Can we still think that reasoning has to do with this kind of faculty that linearly can deduce results from premises? How do we engage with this age of the algorithm where the infrastructure of communication
runs behind or beyond or underneath our very cognition and perception or whether our cognition and perception is informed by algorithms. So the point is how can this be used both to challenge Western metaphysics but also the project of rationalisation but also to rethink reasoning, to rethink subjectivity, to rethink logic with this instrument, rather than just thinking that exactly that your system is based on some kind of accident. And we don't know how
the system works. The question is exactly can we start mapping or unpacking and reinventing or fictionalizing in terms of how is this infrastructure work? And re-inhabiting a future program. You know, if this machine, if the infrastructure, the logical infrastructure of cybernetic computation machine today program the future, preempt the future, this is the argument, you know, whichever statistical calculation you can have you can have on how long you're gonna live, how many accidents you can have in life, or whether you are more likely to do this or that.
All that kind of figuration of your behavior can and must be rethought, taken back. The future doesn't just belong to the branding or the statistical calculation of governance or your health insurance. So how do, the question is how do we collectively develop practices whereby the machine logic can be unpacked to be re-articulated in a different way. So if onto power and accidental sovereignty are models that have exhausted the possibility
of political thinking, because that's what I think they have, they exhausted the possibility of political thinking with machine because they are arguing that these are system of governance of which we don't know the way they operate, or which they operate in such an unstable fragmented way that we cannot decode them, okay? But the question is, my question is, not possibly they can be decoded because the problem of machine thinking is not that they are mysterious, it's that we don't allow for an understanding of how they operate.
We just relate that understanding to a scientific problem. But it's not just a scientific problem, it's a humanities problem, it's an art problem, it's a political problem. It's not just science and economics problem. So how do we do that? How do we re-unpack that kind of relation between reasoning, prediction, patterning, randomness, contingencies unknown? And these are models, you know, we can, there are models there that we can use. So the question is how to rethink philosophy through instrumentality, which can possibly be a good starting point also for
a political re-envisioning of cybernetics rather than a condemnation of cybernetics. Thank you. Automated surgery or automated vehicle results in the loss of life? Who should be to blame? That's a very important question. Insurance companies? Yeah, there is a lot of argument about that because obviously the surgeon, you know, may say that, or the insurance company may say, is the robot that has to be played. There's a fault in the robot. But this is the same as when a drone is shot in a zone area, and instead of shooting a target shoots civilians, yeah?
Who is responsible? Most people are trying to talk about the legalization of co-causality. So it's a co-causality, so not just one responsible, many responsible, so that you unpack all of it. Obviously it has to do with the surgeon, has to do with the clinic, has to do with who made the machine, But also the question is whether the machine can do something that was not programmed for. This is a massive problem because machine learning does allow for decision making. And you have seen, we see this with very simple examples, I mean simple, apparent examples,
with the high frequency trading. where there is a discussion as to, the decision to invest into some kind of a product must be taken so fast, at the nano, nano, nano, nano, nano seconds, that the machine can do much, that there is no time to supervise it. So the decision is taken and something can go completely wrong or can go completely well. You know, so it's a very, it's a very, it's a judicial question and I think must be addressed as a judicial question. And I'm sure that there are court cases in this,
pretty much, yeah. But yeah, it's a very important question. From your argument, can we infer that either now or in the near future, that man will no longer be the primary creative force and that like truck drivers, artists, poets, composers, and writers may become absolutely superfluous? Yeah, that's the question that is called existential risk. There's a group in Oxford that is working on this about existential risk. There is other, also other philosopher like John Serra who says absolutely no, that's absolutely impossible because you know, the famous idea of the Chinese room, you know, you can give the computer all the data by Chinese doesn't
mean he understands the meaning of communication when he communicated, he can just put all stuff together but doesn't understand the meaning, doesn't have the sensibility, the refinement of experience. But this is another content, I guess it's a very old question, because one could look back at the automata of the 17th century and you have this kind of Turkish chess player. I don't know if I have an image, I do have an image. The discussion at the time was, who is this movie? Is it the machine? and it was an example of the highest faculty of reason
that could be embodied by a machine. But something changed in the 1900s that basically this kind of automated intelligence instead became just assigned to clerical work. That's why I got all this picture of women working in the, during the Second World War, you know, encoding, women computers they are called, especially black and white women working together, especially a long history, I think in a kind of very popular reference, hidden figures, the movie in the figures, as we are all talking. So there is this kind of, but they marketed,
according to some historians of cybernetics, they marketed a transformation of automation as high level of intelligence to automation as clerical work. Clerical scientific work, but then you see that the clerical scientific work at NSA becomes a very, you know, a work of encryption, decryption of massive level of complication, calculation, mathematical, so there is a mathematical, not gene, but there is a mathematical rational process going on there that is not just clerical, it's not just task, that's what I mean. But then you have had a lot of experiments in the 60s, 70s with automated art, automated sound, vanguardist
project or conceptual art, vanguardist project. So the question is not about just replacement, but the question is trying to understand what kind of autonomy these learning machines are acquiring and how they are developing a machine aesthetic or a machine sensibility. It's not about replacing but I guess if we don't ask the question will be just a way of replacing it. Like automation is going to hit the job market in less than 20 years on a massive scale globally. Of course this will create different kind of segregation and disequalities amongst workers. The idea
is to try to get a universal income so that labor, working, doesn't exist anymore. So there are, obviously there are a lot of difficulties, but I don't think it's just a matter of replacing. It's a matter of actually enlarging the idea of aesthetic sensibility, reasoning to other forms. And it's not just machine. For me, the machine is just an example, which is my example because I'm interested in it, but it's also an example of understanding other cultures, other modes of reasoning, other modes of sensibility that are not supposed to have a sensibility. So it's about really also looking at the world, and that's why it's important, again, when
I go back to the 80s and the kind of science fiction writing of the 80s, black family science fiction, rather, the 80s, where they were describing modes of knowledge that were developed and sensibility and aesthetics developed already in America in the 80s, where you still have that's a segregated culture. So she was writing about a post-nuclear world. This is Octavia Butler, by the way. But she was describing, she was describing, you know, every day, every day. And so it's about also to, for me, the idea of instrumentality,
recuperating instrumentality is to recuperate the idea that the mean to knowledge is not just a mean. It's not something that you can just use to achieve, like the idea of the colonial empire or capital or Pajart, that you can use an instrument to achieve something. But the instrument develops its own sensibility, its own logic, its own culture. So, you see, on one hand it's the machine, but for me it's bigger than the machine in terms of its political implication. the alliance between animal human machine gender race that was started in the 1980s with Donna Haraway I think is within the circle of critical theory of technology but even bigger in terms of also art expression and architectural expression
is quite important to claim a question? two questions in the back Good evening and thank you for your presentation this evening. It's phenomenal. Not to go off on a tangent or to have a political conversation, but the crux of your argument is looking at this modem from a different perspective. The one question I have for you is the word control you use quite a few times. And if we go back all the way through history, control is not something that is very willingly relinquished. So why? Why then would the existing mechanism allow for it to be questioned to be utilized in a different format?
Yes, control. It's interesting because you can understand control in terms of domination or in terms of self-regulatory system. So in terms of domination, I totally agree that obviously, first of all, it's a struggle and a fight. There's no way, no one is going to give you nothing. for absolutely you know actually you know it's interesting I just been thinking about this one percent question and why we have Brexit and Trump and I'm sorry if I offend anyone because you have to be politically correct is
important but you know if apparently but you know the is this because the domination as basically that kind of domain that kind of maintaining the resourcing such a small megalomaniac way for years, that is neoliberal capital, has created a social degradation, you know, a social degradation to the level that, you know, people thought that actually, at this point, at least I'm talking, I live in London, of course, you know, Brexit is the solution, London didn't vote for Brexit, but you know, is the solution, the solution is completely something that has no meaning whatsoever or has the meaning of actually go back to a conservative nationalist
bread and butter idea but completely dominated by social networks, completely dominated and by algorithmic intelligence and statistical modes of counting on who is gonna vote. It's really, so it's super domination. So that kind of, I don't disagree. I think that computational planetary sovereignty exists, absolutely. The question is, I do not want to relinquish and I don't want to bundle the field of machine thinking to domination. I want to rethink of it in terms of its kind of perspective of moving away from an understanding of power in terms of domination
because the idea of cybernetics control is really based on feedback. It's dialogic, has to do with self-regulation that always has to include the dialogue between the input and the output and can transform in time. So I want to recuperate this also this idea that control can also, or modes of prediction can also be turned towards a constructive alliance, and not just the kind of politics of domination. So I agree with you, but it's a political, and perhaps one could argue theoretical, speculative challenge that I'm just writing science fiction here, but rather write science fiction than not write at all.
And so I don't want to, my challenge is to take away technology from its identification with capital and governance. I want to take it away from that. I want to recuperate because there are examples in history that has been recuperated, both at a conceptual level, a practical level, artistic level. There is a way that one can recuperate, technicity, and when I talk about technology, that is something I'm working on, on technological consciousness. I'm saying that there is a consciousness that doesn't belong to the human or the machine, but there is a consciousness that is including the way the machine sees the world now.
There is something, so during modernity you have all this kind of literature, art, and cinema where technology is your unconscious. It shows that actually something that we don't wanna see because with true technology we have destroyed during the Second World War of course, extermination, colonization, all sorts of violence. It's something that has to be repressed. But all the history, all those stories have come out. We have had feminists, we have had many modes of articulation of class, gender, and race, in terms of political thinking and political action.
So, we can't just make it unconscious again. That is part of conscious, we've not got to claim it back. Something in history has happened, so those period of the 80s, in terms of claiming for other logics, other modes of living. For me, the instrument instantiates needs to be claimed back as something that is part of consciousness, not unconscious, is here. It's happened. So let's stop pretending that it's not real, as it were. OK, I have another question. As the metaphysics of cybernetics starts to disseminate in other areas outside of economics and technology,
and we reconfigure the way we think about things in society, is there a role for cybernetics to commentate on religion? And what's the interaction with religion? What does it have to say about religion? Yeah, that's a very, there is actually, cybernetics already was talking about religion in terms of the kind of, you know, depends obviously what you mean by religion, because obviously there is a kind of question of numerology and religion in terms of whether you understand religion, how do you understand religion? The belief in something that excludes error.
And can't be learned. And there's no belief in that it's all done with that belief. No, the belief that, no, because cybernetics is exactly the opposite, it's based on learning. So, you know, there is no law. There is no law. No, exactly. It's exactly that. It exactly declares the crisis of the law. so that you can't just start from some kind of pure axiom some kind of pure law that is descended on us transmitted on us of the opposite cybernetics exactly by the crib that's what also I think was pointing out by by by bringing the end of metaphysics it means that you can no longer rely on
on truth, but the truth is relative to the environment and the relation and the history in which you are involved. That doesn't mean that there is no truth. It means that the truth emerges from practice, from doing, from learning, from engaging, from struggle. It's not given. But doesn't mean, the question is, when you argue that there is no truth and no law, most critical thinking then has ended up saying that everything is either locally relative or that is impossible to other or relative to your opinion or to your idea to your background or to your community and that has destroyed culture you know even you know for instance
Richard Rorty in terms of localization of culture and that has precluded the possibility of a common project of a common project where you know there is a possibility of arguing for a truth at some point of a logic or an axiom but should not come from something that is not first of all practiced yeah there is a lot of existentialism in this is existential but then it does it jumps after existentialism it needs to jump into the construction of a law it doesn't just stay in existence but yeah good point any other question do you think the new
cognitive intelligence systems are artificial or do you think that they're natural because we teach them based on example they can adapt they can improvise So why? No, I always argue that they are artificial, because naturalization is an enemy for me. Because to go back to nature, they assume that everything, you know, I guess we need a philosophy on nature that accounts for this process of artificialization. But I think even if it's adaptable from the very beginning, you know, cybernetic computation were interested in adaptive system, you know, or system that will tell you, oh, you know, like your Amazon system, they will tell you, oh you like this, I'm sure you like that as well. They kind of preempt what you can learn next,
what you can go next. So Google that writes back to you in case you've been searching for cars. Oh, maybe you'll be interested in this car, and what about this other, this kind of variation, association, in preempting what your desire will be. So preempting your future, that's what I mean by preempting the future. And that's not natural. You know, that is a kind of mode of imposing, imparting a probability of your behavior. It's imparting a conduct. You are being prescribed in behaving in a certain way rather than another. so no I guess naturalization
I think they are always artificial in the sense that there is some kind of relation between a mean and an end there is a reason why this is happening so it's not natural in that way ok thank you very much Thank you.