Artificial Intelligence, Microscopes And Your Next Monday Double Espresso

in LeoFinance26 days ago


We, humans, have a very limited hardware setup. Our sensors are not able to perceive a whole range of data. For instance, we cannot see in infrared, or X-Rays. We cannot hear ultrasounds. We cannot see past microns, or beyond the solar system.

But we are very inventive. Our capacity to create external tools for all the situations where our natural sensors are falling us is incredible. Probably that’s the main difference between us and other mammals, and arguably the biggest driver of our evolution. To see in the depth, we invented microscopes, so now we cannot only see past microns, but we can even witness atoms splitting. To see far beyond the stars, we invented telescopes, so now we cannot only see beyond our solar system, but witness the birth and collapse of stars in incredibly remote galaxies.

The past few hundreds years were amazing from this point of view. But we are approaching a limit. It seems we cannot go further than that, because, well, it loos like it’s nothing more beyond that (or at least nothing that we can conceivably integrate). Of course, we still need to reconcile the relativity theory with the quantum theory, as there are still significant differences in the way world behaves micro versus macro, but we’re getting there.

A New Revolution In Resolution Increase

All the tools above increased our “resolution” in our physical surroundings.

Just like increasing the resolution in our computers displays allowed us to see clearer and clearer images, so was the increasing in resolution when watching everything around us. From micro to macro, we now have gone from a very rough, pixellated image, to a beautiful, virtually endless in depth, sharp and detailed description of our universe.

But, almost hidden behind these visible steps ahead, a new type of resolution is increased these days. It’s a bit surprising, but nevertheless incredibly important. It’s the “resolution” in human behavior.

You see, our range of reactions is not that big. We tend to act in a certain, predictable way, to certain types of stimuli. We even have a name for that, it’s called “personality”. It defines us and we identify with it. This “personality” is the indicator of how we will behave, what choices we will make, what directions we will choose at certain crossroads. But at the core, this “personality” is nothing more than a pairing of stimuli with reactions. A relationship between input data and output data.

You know what’s very good at pairing input data with output data?

Exactly, Artificial Intelligence.

As our computing power became bigger and bigger, and as our collected data increased in size, AI became better and better at predicting. It’s now very good at predicting all sort of things. It can predict prices in markets, quite accurately, it can predict how to go form A to B in autonomous driving, it can predict various types of illnesses, and so on and so forth.

But, at the core, what AI predicts is human behavior.

Somebody Knows You Better Than You Do

Ever wondered how YouTube “knows” how to suggest the next video that you’ll actually like? Or how Spotify creates those playlists so good, that you can’t believe your ears? Behind all this is AI.

An AI so powerful that it now knows you better than you know yourself.

This is a very strong statement, I know. And yet, it’s true.

As you move forward in the digital universe, you leave behind a lot of traces, a lot of data. In 99% of the cases not only you don’t know you’re leaving behind pieces of yourself, but you don’t even care. You “trade” your pieces for some momentarily comfort, or entertainment. On the other side of the screen, though, there are people harvesting this new digital gold, and turn it into the most powerful behavioral microscopes.

So, what exactly that does mean?

Let’s say you love coffee. If you spent, let’s say, between 2 and 3 years on any free social media platform (the biggest data harvesters now) then your coffee behavior is clearer to the owner of those networks than it is to you. They have a bigger “resolution” into that behavior than you. You have a rough, pixellated idea about your own preferences, but they have a beautiful, and scarily detailed map of your own behavior. I can bet they know better than yourself when – and probably where, too – you will have your next double espresso. It may be tomorrow, or it may be next Monday, they will know. You don’t remember every choice you made about coffee, but they do.

Now, a double espresso is not a big deal. But how you vote, well, that is a really big deal. A lot of people will be interested to have a very good image about that. What are your predominant behaviors related to threats (wars, or pandemics), these are also more precious than gold. Why, you asked? Well, because if they know how you react to certain stimuli, they will just trigger that stimuli and reap the benefits. It’s like going to the market to sell something, knowing exactly how much you will sell, for how much money, before the market even starts. They know there will be a market (mostly because they will create it out of thin air), they know you will be there, and they know what you will buy.

Don’t believe me?

I understand. Arguably, this phenomenon is still at the beginning. Yet, all you need is to pay attention. First, there was Cambridge Analytica. Then it was Brexit. Then Trump, then Covid-19. All these events have in common one thing: they all displayed predictable mass behaviors that very few were aware of. All these events where “markets” where you showed up without even knowing you are in the market, and when you bought something without even knowing somebody is selling something. You bought your England back, or you bought your right to make America great again, or you bought your redemption in the face of an invisible health related threat, by rushing to get vaccinated (although this doesn’t stop spreading the infection to others, it just tones down the illness, should you catch the virus).

Know Your Worth

The picture I painted above looks pretty bleak. And I’m not gonna lie, it actually is. As I’m writing this, we’re still in the middle of the biggest behavioral market ever, the pandemics one, and it doesn’t look like it will end soon (although, luckily, there are signs it starts to subside).

But, as bad is it is, there is a way around it.

You see, AI cannot know you without your data. If there’s no data, AI is simply useless, it’s blind. It cannot predict anything. Or, if it tries to predict something, it will go so bonkers that it will be more or less like tossing a coin.

So, you can escape it. The tradeoff is to lower your data footprint.

Now, to a certain extent, this trade between your data and the benefits you get may be profitable. And I mean profitable for you. You may get access to other people, to knowledge, to opportunities.

The real trick is to know when to stop, to know how much (how often, how accurately) you give away your data. And, every now and then, do something completely random.

That will confuse the shit out of the AI.

Image by Joshua Kimsey from Pixabay

Initially published on my blog.

Posted Using LeoFinance Beta


Quite succinctly and to the point. Indeed, all our main panthers have long been studied and described, both individually and in the form of crowd behavior (which is even easier to predict). And as you correctly noted, this may be surprising for us, because we are not machines, we are not able to analyze ourselves in the way that AI does.

It seems to me that if I, as a person, had the ability to focus on self-analysis of my behavior for at least an hour, I would quickly realize that almost all my life, all my decisions were easily predictable and quite typical. Both for my personality and for a representative of the "homo sapiens" species.

We console ourselves with the fact that we are superior to AI due to the effect of surprise - illogical behavior. But let's be honest with ourselves, 99% of our behavior is quite predictable. The most frightening thing is that the more unpredictable things we do, the faster the AI will accumulate statistics on them and predict them by simply lowering their priority.

The most frightening thing is that the more unpredictable things we do, the faster the AI will accumulate statistics on them and predict them by simply lowering their priority.

It doesn't usually work quite like that. Even the slightest modification in inputs can alter outputs significantly. So the threshold between randomness and predictability is still quite fragile, hence, we still have a chance to stay on top of the situation, as long as we manage our data footprint correctly. Including "planting" data outrageously different from what we usually do.

I'm not sure that you can influence them so much. After all, % of non-standard incoming data is estimated. Analytics is based on quantitative analysis. If any analyzing system (including AI) would redraw all the output data by obtaining 1 non-standard package (for example), then this would not work as a whole. It simply puts it off as a pattern with a certain degree of probability of its occurrence in the future.

I'm not saying that you are wrong, this is just my guess :)

As far as I know AI works differently. There are many types of networks, for many specific tasks - image recognition, text predictions (translations), generational adversarial networks, etc.

Very simplified: it takes a whole bunch of input, it sets a desired outcome, then it tries to minimize a cost function, by finding the best next match that will minimize this function. It does this by training the model. This training involves multiple passes over specific neuron layers, back-propagation, gradient descent and quite a few other interesting calculus operations applied. At the end of this, though, there is still a certain threshold that is applied to the results of the cost function. Meaning it's still numbers that you can be on top or below. These numbers are extremely low, 0.00000000000001, but it's still a number that can be affected by the input.

The term AI is misleading, I use it mostly because it's popular, the more appropriate one would be Machine Learning, it's closer to what's actually happening: a machine which learns possible paths, based on parameters.


Posted via

Whoa! Thank you, appreciate it :)

Posted Using LeoFinance Beta

Hive solves this.

Thanks for putting this together. Never put it in those terms but there is a picture being drawn of each of us.

Posted Using LeoFinance Beta

Decentralized social media, Hive included, does solve this, to a certain extent. But what we do is still out there, parseable, data can be churned out of it, etc. Seems to me that we need at least a few layers on top of what we have now, for instance things like privacy and anonymity / pseudonymity. At least we're heading into the right direction, though.

Posted Using LeoFinance Beta

If i catch you at the middle of of your post, that is true an AI is truely blind if there is no data st up, market prediction or analysis may help you to find the statement or statistical data but can't predict the price of a markets coin value, because han are involved here! So AI may help you but can't set up your aim and destination.

Your content has been voted as a part of Encouragement program. Keep up the good work!

Use Ecency daily to boost your growth on platform!

Support Ecency
Vote for Proposal
Delegate HP and earn more