How to protect your attention in a world trying to take it away.

If you’re anything like me, you’re simultaneously sick of and enamored by ‘the Algorithms’. Especially if you have watched The Social Dilemma or read ‘Hype: How Scammers, Grifters, and Con Artists are Taking Over the Internet — and Why We’re Following’.

But if you are in any way suspicious of Google, YouTube, SEO tools, or digital marketing techniques you are aware of the Models. That’s a critical first step, but let’s look at what else we can do.

A model is a simplificaton.

Pure and simple.

The words ‘model’, ‘algorithm’, and even ‘statistics’ sound confusing. But they’re not.

These are all tools to help us understand things that are too big to measure directly.

Models can be very simple. For example.

We want to know who is likely to click on a certain YouTube ad. So we add terms like age, other videos watched recently, time of active viewing, and geographic location. We then collect data for each one of these terms and calculate some inferential statistics like average age, time, and area. Then we use these more focused terms to re-run the model and see how many more clicks we get.

Then we REPEAT and REFINE the model by collecting more data, changing the numerical values of the terms, and adding or removing terms. We continue this indefinitely trying to produce the best results, in this case measured as the greatest number of clicks.

The models themselves can be super basic, consisting of one or two terms. Or incredibly complicated. Think matrix algebra where each term represents a matrix of complex terms.

Simplifications can be complex. But they don’t need to be.

Generally, the best models exhibit PARSIMONY, or contain the fewest terms that provide a high degree of the desired result. Good models are ACCURATE, or provide the desired result (e.g., more clicks). They are also SIMPLE and contain the fewest terms because this makes them less expensive to operate, more easy to understand, and less likely to be distrusted due to being overly complex.

Unfortunately, some models are trusted despite being overly complex. This is an often implemented but rarely critiqued technique. The smoke screen, if you will. Sometimes, overly complex models with low predictive capacity can be used to manipulate consumers

Simplifications probably SHOULDN’T be complex.

You might feel that you’re not smart enough to understand a complicated model. That could be a red flag for a smoke screen. In fact, if you’ll allow me a bit of conspiratorial theorizing, I’d say many times models are over complicated INTENTIONALLY to confuse you into submission. Happens in science all the time. We all it ‘arm waving’. Sort of a, ‘nothing to see here (in the data), so look over here (at my weak explanation).’ Smoke and mirrors.

So what do you do when you encounter a model?

The simpler and more easily understood a model is, the better. If you struggle to make sense of what you are reading, seeing, or being told, be very wary. You can ask for more information, a better interpretation, or alternative explanations. While it’s ok to outright deny the information, I think it is prudent and worthwhile to at least consider you might not understand what the author is communicating.

But beyond the benefit of the doubt, run away. Not all models are useful, remember? And sometimes they are outright manipulative. Learning to know the difference is a key part of being an informed human. This process is not unlike learning to be a scientist. I can’t tell you how to be a healthy skeptic, but it comes from a balance of doubt and an openness to believe.

Learning how to assess information is essentially, faith

Scientists don’t like it when I say that, but there is always an element of faith when making a decision. Because we can never be 100% sure or positive, we eventually have to make a leap into the abyss of the unknown. The questions are, do I believe this? Is there enough information? Are my spidey-senses telling me there is an ulterior motive?

If you answer no to those questions it might be ok to believe. At least for a trial period of further assessment.

You may ask how this differs from simply forming an opinion. Maybe like the flat-earthers or the anti-vaxxers? The difference here is discourse. While you are free to accept or reject a model due to the reasons above, you must also place your position within the context of as many other positions as possible.

For example. Are you the only one who thinks we are actually living in the matrix? Perhaps, for that reason alone, your interpretation is inaccurate. Of course, this can go both ways. Sometimes the majority is wrong.

In the end it is very hard to know the truth. Sometimes it’s simply beyond our comprehension. Other times the competing models are too obfuscative.

In the end, your opinion becomes meaningful when it is informed and practiced. When you consider the evidence, the opposing evidence, and the related evidence, in the context of any biases or persuasive elements you reach a position.

It is also a requirement that this position be FLEXIBLE. To declare the ultimate irrefutable position is to admit your are wrong. If one thing is true in this world it is that things change. Everything changes. It is the one fundamental phenomenon. If you are not willing to change your position, then you are not participating in the discourse.

So How Do We NOT Become Victims?

1. Maintain awareness by assuming your attention is being manipulated.
2. Delay gratification by taking the observer position and asking yourself what is happening.
3. Gather information and combine with intuition to move your opinion into a belief based on facts and feelings.
4. Decide how to proceed.

Algorithmic manipulation of your awareness has been around for a long time, and is only getting better at distracting you. To effectively maneuver through your life it is imperative you master the manipulators before they master you.