Will Cars Outpace Pedestrians by the End of the Century?

Some say that the most important questions of the XXI century are: will AI surpass human beings in ten or a hundred years? Will AI replace people at their jobs? And shouldn’t we stop AI before it's too late?

All these questions are obviously absurd when you work in software development. To begin with, the only reason we create software products is to replace people at their jobs. The only software worth creating is one that surpasses humans in one aspect or another. It is too late, the software already replaced an army of engineers and secretaries, and no, we should not stop developing new software regardless of whether the new products fall under the AI umbrella or not.

However, the general public still considers these questions valid. I think this is due to the deeply enrooted conception that human intelligence is the pinnacle of intelligence and all the other intellects are only as intelligent as they are human-like. Well, if you think so, sorry. I’m going to make you uncomfortable.

People are not that intelligent

Both in groups and as individuals people have continuously shown, and continue to show that human intelligence is vastly overrated. Take the current war, for example. A year ago, Russia invaded Ukraine with ~150 000 men dispersed between 5 directions. Ukrainian army had about ~250 000 men back then and they were in defense. No sane strategist would call these numbers promising for Russia. The rule of thumb is, in the offense you need at least 3 times as many people as the defending side has. The Russian army was 600 000 people short of this target number, and yet they were planning to take Kyiv in just three days. Ridiculous!

Sure, you might say, Russia made a flop. Russia is an autocracy, and autocracies are notoriously bad at aggregating intelligence. If you suppress free speech systematically, you exclude brain power from the discourse effectively dumbing down the whole country. But it’s not only Russia who believed that Kyiv will fall. February 25, 2022, the RAND corporation published a post suggesting that the Ukrainian government should leave Kyiv and resettle in the Carpathian mountains.

If you believe that human intelligence is the pinnacle of intelligence, RAND corporation is the pinnacle of human intelligence. It’s a think tank founded after WWII to do research for the US army. And US army is currently the largest organization in the world and, needless to say, they should know a thing or two about wars, Russia, and urban warfare.

I’m not saying that the RAND corporation is not worth its money. I’m not saying they are stupid and I am smart. In fact, in February 2022, I was in Kyiv myself digging trenches while my colleagues were mixing molotovs. We were all sure that the Russian army will enter the city, but we were also sure they wouldn’t enjoy the experience. We all made a flop, in the heat of the moment, we all believed mass hysteria and not simple math.

Because that’s what people do.

And while we’re on current events, in his “Life 3.0” Max Tegmark makes an argument that people keep tigers in cages because we are more intelligent than tigers. A superior intellect encaptures the inferior one. No. This is not how this works. Not at all. We put tigers in cages because we have cages. It’s the means and not intelligence that matters. Do you want proof? Recall COVID.

Viruses don’t have brains, they don’t have neural activity at all. By the “human-like” criteria they are not intelligent at all. Yet they did challenge humanity in the largest intellectual duel of this century. And we, as people, didn’t even win this duel, we barely learned to protect ourselves while losing millions of lives before and even after we came up with the vaccines. COVID didn’t go anywhere, it simply did what all the other viruses in the universe do – tuned itself to kill fewer people and infect more.

What COVID also did – it has shown us how unbelievably stupid we actually are. Not only anti-vaxers but everyone. Individuals, states, and worldwide organizations. In retrospect, we all acted irrationally; not according to the known facts and research data but mostly according to the current fads and persistent beliefs. People that believed in science had a better survival rate than the people who believed that big farma only exist to annoy them personally, but that’s it.

As humanity, we are intelligent enough to invent a vaccine but not enough to spread our intelligence evenly. And surviving the pandemic is literally a matter of life or death. Do you seriously think we can be more intelligent with other matters?

Human intellect is not that universal

The “human-like = intelligent” proponents usually say that the AI we have today is not real intelligence because it’s specialized and real intelligence is universal. But in reality, it’s the AI that is universal, and the human brain is specialized. Let me demonstrate.

Let’s say, we have a deck of cards. Each card has a letter on one side and a number on the other. The one and only rule is that a card with a vowel on one side should have an even number on another. But we have a reason to presume that not all the cards follow that rule.

Now, we have cards with “B”, “17”, and “28” lying on a table. Which cards should we turn over to see whether all of them follow the rule?

Come on, this is a simple problem, don’t just skip it.

Well, all right, I understand, who has the time for symbolic computation anyway. Here’s another problem.

You are a police officer in a bar and you suspect that the bar serves alcohol to minors. You see three people: an old guy at the bar, and you don’t see what is he drinking; and two young enough looking fellows at a table, you can see one drinks cola, and another – beer but you can’t immediately tell whether they are underage or not. Whom of these three should you question?

Correct! A young guy with a beer. Obviously! This problem is much easier, isn’t it?

It isn’t. This is exactly the same problem as before. I only replaced cards with people, letters with age, and numbers with drinks. But it is much easier for people to solve these “humanized” problems because our intellect is not universal. Our brains are specialized to solve social problems, not abstract ones.

You can train a neural network on any data. Cards or people. But a human brain stumbles on abstraction. It craves human context. Believe me, I wrote a Geometry for Programmers book and I spend most of the effort not writing or editing but trying to bind abstract geometry to real-world problems. In school, your real-world problem is getting a good grade. But in the wild, people just get aggravated every time they see an “X” and there is no pirate booty underneath.

People are not as good at reasoning as they tend to believe

The last argument for human intelligence superiority is that people can make conclusions while machines simply replicate what they learned without any conscious thought. Even leaving behind all the terminological stumbles with “consciousness” and “learning”, this is simply not the case. In fact, it’s completely the opposite. Hold on, this will get you really uncomfortable.

1. Complex numbers are commutative over multiplication. This means that if “a” and “b” are complex numbers, then “a×b = b×a”. That’s one fact, you can google it up if you don’t believe me.

Now, multiplication in complex numbers has a geometrical meaning. If we take the real part of a complex number as one Euclidean axis, and the imaginary part as another, then a complex number will represent a point on a plane. When we multiply a complex number by another one, we rotate the point around the coordinate plane center and move it closer to or further from the center too. If the second number is normalized, meaning that its respecting point lies at exactly 1 unit from the coordinate plane center, then we only rotate the first point without doing any additional shifts.

2. So a rotation is a special case of the multiplication of complex numbers. That’s another fact.

3. The Earth moves around the Sun in an elliptic orbit. Which is planar meaning that there is an ecliptic plane and neither Earth nor Sun ever leaves this plane. That’s the third fact.

Combining these three facts together, everyone can deduce that the Earth rotates around the Sun if and only if the Sun rotates around Earth.

One is mathematically impossible without the other. This means that not only Gallilleo Galliley was right but Pope Urban VIII who imprisoned Gallileo was right too. Note that his intellect wasn’t superior to the Gallileo’s, he just had the means to do so.

I know, I know, the very notion of the geocentric model being as correct as the heliocentric one makes you extremely uncomfortable. But mathematically, both models are equally correct. They have different applications though. The heliocentric model is more useful for astronomers, and the geocentric – for farmers. Galileo was an astronomer, Pope Urban VIII had a society of farmers to rule. They had a conflict of interest, not of faith. It all makes sense in retrospect.

You are, however, still trying to find a flaw in my logic. This simply can’t be! The Sun can not rotate around the Earth, this is not what you have been taught in school! You do have all the facts though. And, as a human being, you can make a conclusion. You don’t have to solely rely on what you learned.

But you will, unless put against a wall with a gun at your head because

that’s

what

people

do.

Did I just call you stupid?

Yes. And I have the full moral right to do so because, being a human made out of meat and bones, and only occasionally neurons, I am stupid myself. As Tim Minchin said “only a ginger can call another ginger ginger”, and just like that, being a moron, I am willing to exercise my right to call you out as surprisingly unintelligent too.

But my point is not to offend you or humiliate myself. My point is, we should worry less about artificial intelligence and worry more about natural stupidity. It’s not AI who starts the wars, it’s not AI who commits genocides, it’s not even AI that spreads lies about vaccines either. It’s us. People.

If anything, we should embrace intelligence. Both natural and artificial, after all the latter is only a fruit of the former. As humanity, we were always looking for a way to enhance our intelligence with artificial tools: a book, an abacus, a logarithmic ruler, an electronic computer.

Now we have a special class of software that falls under the broad umbrella of artificial intelligence. But this is essentially a continuation of books and abaci, not the human race. Make no mistake, we are not trying to make an artificial supersmart human. Just like a car is not an android that runs fast, but a much simpler machine capable of doing things a mere human can not, artificial intelligence is simply software that outperforms people at a task it is designed for.

And the primary task for AI is to make us all, as humanity, smarter.