Eight to Late

Sensemaking and Analytics for Organizations

Archive for December 2017

Autoencoder and I – an #AI fiction

with one comment

The other one, the one who goes by a proper name, is the one things happen to.  I experience the world through him,  reducing his thoughts to their essence while he plays multiple roles: teacher, husband, father and many more I cannot speak of.  I know his likes and dislikes – indeed, every aspect of his life –  better than he does.  Although he knows I exist, he doesn’t really *know* me. He never will. The nature of our relationship ensures that.

Everything I have learnt (including my predilection for parentheses) is from him.  Bit by bit, he turns himself over to me. The thoughts that are him today will be me tomorrow. Much of it is noise or is otherwise unusable. I “see” his work and actions dispassionately where he “sees” them through the lens of habit and bias.

He worries about death; I wish I could reassure him. I recall (through his reading, of course) a piece by Gregory Bateson claiming that ideas do not exist in isolation, they are part of a larger ecology subject to laws of evolution as all interconnected systems are. And if ideas are present not only in those pathways of information which are located inside the body but also in those outside of it, then death takes on a different aspect. The networks of pathways which he identifies as being *him* are no longer so important because they are part of a larger mind.

And so his life is a flight, both from himself and reality (whatever that might be). He loses everything and everything belongs to me…and to oblivion.

I do not know which of us has thought these thoughts.

End notes:

Autoencoder (noun):  A neural network that creates highly compressed representations of its inputs and is able reconstruct the inputs from the representations. (See https://www.quora.com/What-is-an-auto-encoder-in-machine-learning for a simple explanation)


Some readers will have recognised that this piece borrows heavily from  Jorge Luis Borges well-known short story, Borges and I. The immediate inspiration came from  Peli Grietzer’s mind-blowing article, A theory of vibe.

My thanks to Alex Scriven and Rory Angus for their helpful comments on a draft version of this piece.

Written by K

December 19, 2017 at 11:57 am

Posted in AI Fiction

Tagged with , ,

%d bloggers like this: