Artificial Stupidity (v3)

I made a mistake. Irrationality isn’t caused by multiple needs pulling in different directions. I’d say that this is the pattern of irrationality … but it is not the cause of it.

If I need to put some petrol in the car and at the same time I need to get to work as quickly as possible … then I don’t become irrational. I estimate the chances that I have enough petrol and if I think I dont then I go to the petrol station and cut my loses. I dont get caught in an internal conflict… at least not for long.

That pattern where things get pulled in different directions without any apparent reconciliation … or at least no permanent reconciliation … that pattern of irrationality, where things that are very clear to one person, but so utterly confused or not even perceivable by the conflicted person … that pattern, that ‘stupidity’ ….. is what I would like to recreate, replicate, simulate, model, generalise in a programmatic structure.

As shown in the example above, just because two needs conflict does not mean they cannot be reconciled. I need to identify the kinds of needs that are not reconcilable. What makes a need unreconcilable?

I suspect that the creation of the need that is unreconcilable will involve some level of reflection. I’m talking about software reflection. For example, there will be a need which will say “I must no longer have any needs”. And the problem will involve some such thing where the need itself can never be satisfied by the simple reality that it exists. The need to “no longer have any needs” is denied satisfaction by its own existence. Its like some sort of self-referential infinite loop… that gets occasionally triggered then released. It does get released.

… its something like that. And that kind of structure is quite easy to model programmatically. But I haven’t got my finger exactly on it yet.

This entry was posted in artificial stupidity. Bookmark the permalink.

4 Responses to Artificial Stupidity (v3)

  1. mark says:

    I get the feeling that this problem might benefit from a less high-level ‘modelling’ approach than the type you’re developing (in the previous post, but also in your comments here). In the previous post you say “What I’m not convinced by is that needs are processed in some order. I wonder if perhaps there are some needs which have a natural greater priority.” I’m also not convinced that needs are processed in some order, and my intuition is that your sequential model limits how interesting any results will be.

    Martin’s comment regarding Maslow’s hierarchy of needs is partly an answer to your final question regarding a natural prioritisation of needs — you could include these as weights to your sequential model, but it won’t change much other than the ordering of needs, which you already incorporate.

    Another reason you won’t get interesting results from the current model is that you have no legitimately interesting source of new observations; as long as you (the programmer) provides the observations the outcome of the algorithm is 100% predictable. You need some way of actually generating observations.

    An answer to both the sequential model and the new observations problem might be to parallelise your algorithm, to mimic the massively parallel behaviour of our brains. A simple way to do this would be to define a Person object, which has six categories of needs (from Maslow), and assigns a different weighting to each category. Instantiate several Persons and define a universal goal, then let it run generation-after-generation and see what happens. A bottom-up modelling.

    One possible scenario might be this: 10 Persons in a community, and they decide that they want to build a house together. Each person has different motivations for building a house. People who weight security highly want it for security (I.e. want to see it through to the end). People who weight socialisation highly but also weight immediate biological comfort highly want to be involved with the actual building because it’s a social activity, but might get sick of working after a while. People who weight social standing highly want to live in the house after it’s built because it’s a prestigious setting, but might not want to work very much to achieve it. Self-actualisers are just happy to keep working on it for the sake of progress, and so on. Also: if enough Persons decide a Person isn’t working enough, they get kicked out entirely.

    Then, you have an iterative process: at each generation, every Person has the chance to lay a brick. Each person knows how much each other person has done (forming part of their observation), but each laying of a brick incurs a biological cost (forming another part of their observation). Then, given all this information (how many bricks were laid in the past generation, who isn’t working hard enough, a Person’s internal state (i.e. tiredness [i.e. bricks laid in past generations]) they make a decision whether or not to lay a brick in the next generation.

    This is reasonably similar to how neural networks work. Each neuron takes a set of weighted inputs and if they sum greater than some threshold, the neuron fires otherwise it doesn’t. If you track the laying rates, any Persons that get kicked out, I think you’ll get some interesting outcomes, and then you can tweak the Person weightings and run the whole thing again. I imagine you’ll get a nice picture of the type of balance required to achieve a goal in a parallel environment of conflicting needs.

    • Hi Mark,

      I do realise that the brain is probably massively parallel… and that my modelling does not reflect this at all. However! … I’m not attempting to model how the brain thinks. I’m only attempting to model the part of the brain that causes irrationality. The rest, I dont care.

      Ofcourse, I have a pre-conception about what makes us irrational … and what I am essentially doing is attempting to validate that pre-conception by articulating it … through writing the blog and through the code abstraction.

      I dont think it is the conflicting needs that causes the irrationality, I think it is only a specific type of need … which cant resolve itself. But to demonstrate this … to articulate it, I need to identify what those needs are and explain why they cant resolve themselves.

  2. mark says:

    I wasn’t invoking massively parallel for the fun of it. You have entered into the class of problem that is actually properly philosophical (rather than artists thinking they are philosophers). You intuit you have some pre-conception that is for some reason inarticulable. It is inarticulable because every time you start the sentence in your brain the second half of it doesn’t make sense. This is not a logical failing of your brain, it is a product of indeterminism and complexity. That’s why I suggested you try something from the ground up than look for a high-level solution.

    • Hi Mark,

      My pre-conception is perhaps better described as a perception. As an artist, I am trying to re-create that perception in others via a work of art… Perhaps the blog is the work of art … perhaps not. It is definitely philosophish but … whatever, I dont care what it is. Its a perception, and I am trying to see if I can insight it in others.

      I havn’t articulated it explicitly in words. I could do that. It would only take several sentences. But to do so would be like taking a photo of a beautiful hill in the summer sun on a blue sky day …. it would just make you think of a Windows 98 desktop. So I am weaving around it attacking it from different sides. Trying to see if there is a way of describing it that incites nothing but the essences of what it is.

      Your comment there actually establishes your own insight … does it not? That irrationality is the result of complexity and indeterminism? Do you think this is correct?

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s