Every thought you have strengthens those synapses at the expense of others. How does this black box problem make it difficult to learn about learning?

You Are What You Think

You Are What You Think


July 13, 2020

Author: Victoria

I have a question about how we learn about learning. Here’s a thought experiment to explore it.

Thought Experiment: Comparing Hardware

Imagine a dozen computers, each running a different set of programs. We expose these computers to the same information, each recording with similar sensors, with only minor manufacturing defects modifying the experience. We then test the computers interpretation of that data with a series of questions and find very different outputs from each. We draw conclusions about the relative power of these computers hardware based on these results. We decide that computer 3 must have the fastest RAM, that computer 8 has a superior storage capacity and computer 9 appears to be lacking in either.

Differentiating between hardware and software

Assuming we cannot simply crack open these computers and check our conclusions, how would we know they’re right? We can’t. Without knowing how their respective programs interpreted the information, how it was transformed, where it was stored and how frequently it was retrieved, we cannot draw any meaningful conclusions about the computers hardware from output dependent upon both hardware and software. How could we determine to what degree good output was the result of the speed of the hardware, and not the quality of the code it was running?

Differentiating between the brain and the behavior

Since we can easily crack open computers and compare their contents, this is not about computers. This analogy is about people. Comparing people is a very important daily activity for society, yet it is a faulty one. We use heuristics and tests to rank black boxes that we do not understand.

If the test giver has some notion of what best means, then much of this complexity isn’t relevant. It does not matter how the brain used the information, it only matters that they know it (assuming creation isn’t required //todo). However, If the test giver wants to understand the brain itself, both hardware and software do matter.

The Infinity of Thought

Anyone who has day dreamed, knows it is possible to be physically present, but mentally absent.

‘two things are infinite, the universe and human stupidity’ -Albert Einstein

As stupidity is a subset of the entire imagination, we can take it that this wise guy also thought the imagination was infinite. Exactly where one dwells in this infinity of thought may be invisible to the outside observer, and below consciousness to the inward observer, but it is the cause to the effect we’re interested in.

The probability of thought

A thought has a 'substrate', a physical mechanism that underlies its existence. Thoughts are electrical currents across the brain, triggering action potentials, the release of neurotransmitters, strengthening the neural connections with each access. While each person might have physically been in class for two hours, the amount of time spent thinking about the course material is individual. This is compounded outside of the classroom. Some dutifully do their homework, the highly motivated may even contemplate the lesson in their own time. The value of this recurrent thought in unorthodox settings is even greater, considering the number of creations that have emerged during seemingly unrelated activities (todo: write article about inventions at unlikely times). Other students feel their duty is done when the class period is over. They apply minimal effort to their homework, if any at all, and they spend their time dwelling on other facets of the universe.

All people exist on a spectrum between thinking a thought all the time, and none of the time. How can we compare two thinkers without knowing where on this spectrum they fall? What motivates them to think about the topic? Are they a fast learner who spent little time contending with the subject? Are they a slow learner who spent a great deal of time enjoying the subject? These combinations (todo: write article about combinations of behaviours) effect the kind of insights such brains might produce in years to come.

An employer can sidestep much of this complexity, as a candidates desire to imbibe new information is itself a desirable and measurable property. Anyone who wishes to understand the brain itself, must take this complexity into account.

What do we want to think about?

Life is thought. Every thought strengthens some electrical network at the expense of others, making our brains more unique over time. So what do our respective hobbies that we choose to fill our heads with teach us?

(todo: write articles looking at common hobbies lessons)

One product, Two Unknowns

Without the ability to measure how the information of interest has been used or transformed, any comparisons of brains or computers is moot. That which we cannot measure is deemed necessarily useless to us, even if it is the motor driving the machine we study. Any conclusions we might try to draw depend upon our own ignorance of intelligence and how our brains work.

We often speak of intelligence in terms of raw processing power of our fleshy equivalents of RAM or threads, but incongruously we know that it is what we choose to think about that drives our action potentials. If how we spend our time wasn’t the critical factor, would we invest so much into educational material for children, when the outcome is fixed by hardware anyway? Intelligence is a result of the software and the hardware. Only when we have processors running the same program, can we actually judge their hardware's potential. A powerful computer running pointless programs will only compute pointless answers more quickly.

References