Why we do anything at all

Is consciousness the instigator of all action, or is something else at play?

Imagine this..

The year is 2035 and you can now buy a new type of computer chip that mimics neurons in the brain. These chips are designed in a way that enables you to click them together in various combinations; rather like Lego.

What’s beautiful about this is that rather than performing machine learning by “simulating” networks of neurons via software on a computer, you are instead building “real” networks to learn specific things.

Why does this matter?

Well, any real object is governed by the physics of the real world, rather than an abstraction like we see in computer software. If you think about it, all software does is convert everything into 1’s and 0’s - very far from how the real world works.

As Christof Koch puts it “just because you can simulate weather on a computer, it doesn’t mean you will actually create a storm”, or something along this line.

I can’t stress the importance of this point enough - I don’t think it spoken about enough in the context of AI either. You even hear of so called “experts” who claim that a software based AI may have some kind of consciousness. It is disappointing that they miss this really important distinction about simulated v real analog-based systems. Anyway, I am digressing..

---

The purpose of my entry today is to discuss a recent thought I had while writing code for optimising batteries and Virtual Power Plants (my day job is in the energy sector). With any optimisation code, you have some objective, and you tell the program to work towards this objective. For example, you may wish to maximise revenue from a battery, or you may wish to minimise how much energy you import from the grid to your home, and so on. The key take away here is that the code now has an objective - something to work towards.

So naturally it got me thinking…do our own brains have an objective function of some sort?

Think about my example from earlier for a moment. You have built a real network of neuron chips to form an artificial brain of sorts. You want to create a creature similar to a worm that simply moves by squirming around (weird choice, I know). So you train your newly created artificial brain on the behaviour of worms to try and replicate this in your machine you are creating. After months, you have finished the body, the brain, and all training required.

You proceed to place your new machine on the ground and …

Nothing.

The robot, even with a brain that uses real chips, sits lifeless on the ground. Your mind starts racing around. How could the robot instinctively know what to do without having to preprogram anything into it?

In other words, how can we design a robot in a way that it becomes ‘alive’ - where the robot is able to continuously create subconscious objective functions for itself to fulfill?

I believe the answer to this is deep-rooted in the question of consciousness. Perhaps consciousness itself is the impetus needed for any sentient organism to do anything at all.

Is there another step in-between consciousness that guides the conscious experience to do something? Or is it that consciousness itself is the driver?

---

For the record, worms have just 300 neurons! That’s all it takes to create a creature which is alive and moving, without following a pre-set script or pre-defined objective functions.

I would argue that a worm MUST have some level of conscious experience to do this, but even then it still begs the question of how does consciousness translate into action?

I’ll leave this with you.