• 0 Posts
  • 5 Comments
Joined 1 year ago
cake
Cake day: October 26th, 2025

help-circle
  • Thanks for this!

    I very much agree that I have over-simplified here, and likely mystified things a lot. I was only trying to point out that assuming our brains work like computers with lots of neurons and processing power implies an assumption that consciousness is algorithmic in it’s process. An assumption that consciousness is non-algorithmic, in that there is no way to demonstrate that outcomes are a direct result of a set of specific logical steps, is equally as valid an assumption to make about consciousness, I think.


  • I just mean that consciousness is obviously a complicated process that requires a sufficiently complicated material system to operate.

    This is the assumption that you are making, yes. You are assuming consciousness is a complicated process that requires a complicated material system.

    We only KNOW, with certainty, about consciousness in humans, who probably-not-coincidentally have the most capacity for biological computation.

    Here you are assuming that since humans have the most capacity for biological computation, and humans are the only ones we can say with certainty are conscious, than consciousness therefore requires a lot biological computation, or material processing to function. This is the assumption that consciousness is algorithmic. The term ‘computation’ by definition refers to an algorithmic process. ‘Compute’ = run the input data through the algorithm to produce the output.

    This is a perfectly valid assumption to make. All I am trying to point out is that you are making the assumption that consciousness is algorithmic and requires material processing to work. No one knows whether that assumption is right or not, that is why it is called the ‘hard problem’. An other assumption to make in approaching the hard problem is to assume that consciousness is NOT algorithmic, and therefore does not necessarily require an excess of material processing for it to operate. In this case organisms without a ton of biological processing capacity could still experience consciousness.

    This is a good example of idealism, as opposed to materialism.

    I mean, sure? But really, this is satire. The point is that in 1726 the eggheads at the Royal Society argued that following only pure logic and the laws of physics is the pinnacle of how societies should be run. A technocracy. Swift is satirizing that argument by demonstrating exactly all of those things you just said about creating waste and no new information or whatever.

    The point I am trying to make with that example is that LLMs function exactly how Swift’s Engine does. The idea behind LLMs and genAI is that if we input enough data, and iterate enough times for training, eventually the LLM will know everything and be able to do everything, and so all we have to do is prompt it and our problems are solved, no one in the world ever needs to work again, we’ve solved the climate crisis, etc. Swift’s Engine is exactly how we train our LLMs today, and as you’ve pointed out, it’s not a very efficient way of doing things. I just think it is fucking cool that 300 years ago a guy came up with such a brutal burn that it is still relevant and valid in our ‘technologically advanced’ society today.


  • From the context of your argument I am interpreting the ‘processing power’ you are talking about as the physical power or energy needed to perform physical work as input data is being physically manipulated by a series of logic gates per the algorithm.

    Perhaps I am misinterpreting what you are getting at when you mean ‘processing power’. The distinction I am trying to point out is that there is no way for any algorithm to be able to generate a true random number. All of the energy in the universe converted into processing power will still never be able to produce a true random number as an output through an algorithm.

    I am also likely misinterpreting what you mean by ‘material’. Is the number 3 a material thing? We can use materials to interpret, or represent the number 3 in the physical world, but the number 3 itself does not have any material basis in our physical world, it is just an abstraction created in our brains. If our physical brains do not exist to be able to create that physical abstraction, does the number 3 even exist at all, does it have a material basis?

    I do not mean to argue about this at all, there is no right or wrong here. I’m trying to point out that some people assume consciousness is algorithmic, some people assume it isn’t. People who assume consciousness is algorithmic are the same people who believe that the genAI that we have today will eventually become conscious, or sentient, once there are enough data centers to be able to provide enough processing power to eventually reach the level of consciousness we are talking about.

    People who assume consciousness is non-algorithmic genuinely believe that the genAI we have today will never actually be conscious, because all our AI is today is a bottom-up algorithm iterated billions and billions of times using all of the data that exists in our world. However, if consciousness really is non-algorithmic, just like the case of more processing power never being able to produce a random output, more data centers will never able to produce consciousness, or AGI.

    Roger Penrose’s 1993 book ‘Shadows of the Mind’ is a great resource to see the distinction. There Penrose very thoroughly distinguishes between top-down and bottom-up algorithms, and the difference between algorithmic and non-algorithmic processes. His description from 1993 of what today we call ‘deepfakes’, for example, is incredibly interesting.

    Again from the algorithmic standpoint, I think my favourite description of how our genAI works today comes from Jonathan Swift’s ‘Gulliver’s Travels’ from 1726. In part 3 of the tale we are introduced to a technocracy and a description of a device called the ‘Engine’ in the Academy of Lagado. This machine is made up of all of the words in their language written out on a giant grid. There are a series of rotating levers that changes the order of all of the words on the grid. The argument is that if you rotate the levers enough times the words will eventually arrange themselves into logical sentences that say real things. If you write down all of the logical sentences that the machine creates, and if you keep rotating the levers billions and billions of times across thousands of different Engines writing down every logical sentence that comes up, eventually you will have written down everything that is knowable in their universe.


  • Algorithmic processes produce an output based on the input. Something gets input into the machine, like visual stimulation for example, the machine takes that input and performs a series of logical operations (the algorithm) on the data and produces a specific output. In this example when we see the colour blue (the outcome), our eyes take in as input the wavelength, frequency, intensity, etc. of the light we are seeing and our brains take that data, run it through logical steps like comparing it to past light frequency data we have in memory, or sorting it along a spectrum of different light frequencies that we are also seeing at the same time or whatever. These type of algorithmic processes are completely deterministic, as in the output can be determined based on the features of the input data and how that data gets manipulated through the algorithm’s logical steps.

    Non-algorithmic processes occur when there is no set of logical steps that will determine the outcome based on the input data. Quantum mechanics, for example, is a stochastic process where the outcome of a measurement cannot be determined by the input parameters, but rather the outcome is determined by a set of statistical probabilities of possible outcomes. Of course the Schrodinger equation is deterministic and algorithmic, however when an outcome is measured that determinism collapses along with the wave function and the measurement becomes stochastic in nature. The collapsing of the wave function is a non-algorithmic process and there is no consensus among physicists on how that ‘collapse’ process actually works in nature, hence the plethora of different interpretations of quantum mechanics out there today.