The Distribution of You

Posted on May 19, 2020 by Hoagy

Something that's on my mind: the multi(re)verse, that which is the total distribution of what ground truth is likely to be (have been?) given the sum total of preserved information. In a future where our human ambitions are totally usurped, this is a core object of foundational value.

What this means is, let us imagine all of the information we have on a person in history. We may know their date of birth and some basic facts about their life.

At the moment this is essentially equivalent to our indivudual mental representations of history but in time it could come to be a fully realised object whereby simulations are judged according to their fit with the information available to the present. Perhaps this object will be constantly evolving into the future, and this should be true if there is constantly new information being discovered. On the other hand, if there's trust in the information saving procedures and no malicious cation then it should stay stable.

It also differs importantly in that history is the attempt to recreate a single true version of events but with the ability to simulate worlds the distribution becomes an object of value in its own right.

The is also the smaller but equally interesting equivalent for a single person: what is the distribution of the mental and physical properties which a past individual had? Many people, especially those interested in cryonics, talk about information death: the point at which the information about the way that one's brain is wired up is lost forever. Under this view, information death is still a loss of information but a blurring, rather than a deletion of the knowledge required to reconstruct a person.

Perhaps we should try and save the parts of ourselves that we care about being replicated and nothing else, to maximise the variation in those dimensions that are less important to us. Is it better to be well defined? If we are at such a point in history

I suppose we shouldn't assume even with all the simulating power of galaxy-scale quantum computing that this is a truly tractable problem, perhaps there could always be effort to generating new possible consistent trajectories. Would we have the capacity to save something so complicated? I suppose the compression of the computation of the world and the compression of the saved state are probably quite closely linked, if we can do one we are likely to be able to do the other.

I wonder if we will value access or simply existence? If we can run a quantum algorithm such that all possible consistent histoies are simulated, but wecan only access a single one, would this be sufficient? I think not since this is ultimately a pretty frivolous exercise in human nostalgia and we would want to be able to explore the data - that these minds have come into existence would probably not be so important if we spend a different fraction of our energy in exploring mind space.

Makes me much more conscious of the way in which I am becoming sensitive to reality by walking around. I wonder how many of the memories that I will never recall again will be recoverable in theory. I suppose like everything, vague and half overwritten memories will be recoverable only probabilistically, and perhaps combined with others' to determine history.

How much of the data on those in living memory is saved in the form of person rather than in things? It seems that in the modern world the production of data probably outstrips our memories in bits, if we really had the ability to reverse engineer events from data. What about the internal lives of people? For this human memories are in a far more amenable form but it seems plausible that the data eventually becomes more restrictive on history. Perhaps they encode very different things, written words showing the nature of the internal connection but things like emotional reaction and overall. I wonder if those studying history through literature would expect that huge amounts can be recovered in theory, though they would bea biased sample.