Something which struck me from yesterday's post was the whole business of 'deletion' with technology. The absenting of something through destruction (like a fire) usually leaves some kind of trace. Although one might inspect the magnetic bits on a disk after a file has been erased (and sometimes recover it), the whole principle of a computer is the obliteration of things without trace. Indeed, erasure is more fundamental to computing than storage: it is the principle of erasure that makes storage possible. It may be to do with the binary nature of the computer, the flipping of bits - but with a bit-flip, the 1 that was there is absolutely gone. Not "crossed-out", not smeared with an eraser, but completely gone. Might it be because of the complete obliteration of things that its opposite, the complete permanence of things, also becomes possible. These are the fundamental principles of 'data'.
Animal experience is not naturally attuned to complete obliteration. Removal or absenting, always leaves a trace. It is through the trace that the passing of time can be marked, that the passage of events can be coordinated with. In the passing of events, and the destruction of things, we read the processes of each other, learn more about each other, and gradually learn to coordinate ourselves better with each other. In the computer, we have to remake this kind of passing of time through using storage, maintaining state, and presenting a consistent interface. In tracing the obliterations performed by others, we have to think in the way the machine 'thinks': we have to consider the behaviour at the machine that the person deleting something might have exhibited. But we know that this behaviour at the machine is something only somebody familiar with the machine would understand. Whilst the language of obliteration in the physical and animal worlds is a language immediately comprehensible to any other member of the species (think of a death), the language of the obliteration of data in a computer is a language comprehensible to those who know the software. We now live in a rarified and particular universe. It is remarkable how we have become accustomed to something so fundamental and yet so alien to our nature.
One of the major advances in computer technology is going to involve a transformation of this situation. I expect that the interfaces of the future will allow for the presentation of decay and "erasure with traces" in ways which we only yet know in the physical domain. When this is achieved, we will think about technology differently. Technology today is stressful - it leads us to commit to practices where we often feel hidebound to comply with (indeed, the managerial control of technology has largely turned it into a tool of corporate compliance!). The working-out, the process, the traces of thought, the sheer messiness is all gone. All we get are clean and sterilised forms to fill in. This is partly because we have a particular view on the causality of information: that decisions can be taken as a positive consequence of "being informed" (so the information is seen as a kind of billiard ball bouncing into a 'decision' ball).
But information isn't like this. What happens is more processual, and the causal relationship is one which involves the discovery of the nature of colleagues and others and their thinking. Decisions are forms of agency taken in the light of the likely impact on those who are important to us, and who we hope to know something about. Information causes decision through a negative process of constraint, not a positive process of deduction.
What is fascinating in this however, is that the future computers which do not erase, but allow for decay and "traces of destruction", may not even operate in a conventional way. With no erasure, what is also reconfigured is the very idea of storage. When storage and programs are separable entities in a Von Neumann architecture, storage and programs are one and the same structure. The kind of sophisticated weighted logics that we see in Neural Networks may well replace the binary logic of current technologies (this is not to say that neural networks will be the model - I have some reservations about that - but that the simple on-off will be replaced, possibly by something more analogue.)
More fascinating is the fact that the machine with decay is almost 'life-like'. This is not to say that this is AI. But it may be a kind of artificial life with which we become attached; not something which we simply "use". At the root of the attachment process will be greater coherence with primeval instincts for following the traces of destruction of a thing.
The computers we have today are like strange phantoms who fascinate us because they appear suddenly without warning and disappear equally suddenly. We have learnt a meta-language to describe their behaviour. But it is odd behaviour. Imagine if human beings were like that: that on dying, for example, people simply disappeared; There was no dead body to grieve over. It is hard to conceive of how we would relate to one another. But love and attachment would surely be changed in these circumstances.
But might we love the computers of tomorrow?
Animal experience is not naturally attuned to complete obliteration. Removal or absenting, always leaves a trace. It is through the trace that the passing of time can be marked, that the passage of events can be coordinated with. In the passing of events, and the destruction of things, we read the processes of each other, learn more about each other, and gradually learn to coordinate ourselves better with each other. In the computer, we have to remake this kind of passing of time through using storage, maintaining state, and presenting a consistent interface. In tracing the obliterations performed by others, we have to think in the way the machine 'thinks': we have to consider the behaviour at the machine that the person deleting something might have exhibited. But we know that this behaviour at the machine is something only somebody familiar with the machine would understand. Whilst the language of obliteration in the physical and animal worlds is a language immediately comprehensible to any other member of the species (think of a death), the language of the obliteration of data in a computer is a language comprehensible to those who know the software. We now live in a rarified and particular universe. It is remarkable how we have become accustomed to something so fundamental and yet so alien to our nature.
One of the major advances in computer technology is going to involve a transformation of this situation. I expect that the interfaces of the future will allow for the presentation of decay and "erasure with traces" in ways which we only yet know in the physical domain. When this is achieved, we will think about technology differently. Technology today is stressful - it leads us to commit to practices where we often feel hidebound to comply with (indeed, the managerial control of technology has largely turned it into a tool of corporate compliance!). The working-out, the process, the traces of thought, the sheer messiness is all gone. All we get are clean and sterilised forms to fill in. This is partly because we have a particular view on the causality of information: that decisions can be taken as a positive consequence of "being informed" (so the information is seen as a kind of billiard ball bouncing into a 'decision' ball).
But information isn't like this. What happens is more processual, and the causal relationship is one which involves the discovery of the nature of colleagues and others and their thinking. Decisions are forms of agency taken in the light of the likely impact on those who are important to us, and who we hope to know something about. Information causes decision through a negative process of constraint, not a positive process of deduction.
What is fascinating in this however, is that the future computers which do not erase, but allow for decay and "traces of destruction", may not even operate in a conventional way. With no erasure, what is also reconfigured is the very idea of storage. When storage and programs are separable entities in a Von Neumann architecture, storage and programs are one and the same structure. The kind of sophisticated weighted logics that we see in Neural Networks may well replace the binary logic of current technologies (this is not to say that neural networks will be the model - I have some reservations about that - but that the simple on-off will be replaced, possibly by something more analogue.)
More fascinating is the fact that the machine with decay is almost 'life-like'. This is not to say that this is AI. But it may be a kind of artificial life with which we become attached; not something which we simply "use". At the root of the attachment process will be greater coherence with primeval instincts for following the traces of destruction of a thing.
The computers we have today are like strange phantoms who fascinate us because they appear suddenly without warning and disappear equally suddenly. We have learnt a meta-language to describe their behaviour. But it is odd behaviour. Imagine if human beings were like that: that on dying, for example, people simply disappeared; There was no dead body to grieve over. It is hard to conceive of how we would relate to one another. But love and attachment would surely be changed in these circumstances.
But might we love the computers of tomorrow?
No comments:
Post a Comment