The limits of AI…

Photo by Spiros Kakos from Pexels

A team of mathematicians and AI researchers discovered that despite the seemingly boundless potential of machine learning, even the cleverest algorithms are nonetheless bound by the constraints of mathematics.

Research showed that a machine’s ability to actually learn – called learnability – can be constrained by mathematics that is unprovable. In other words, it’s basically giving an AI an undecidable problem, something that’s impossible for an algorithm to solve with a true-or-false response.

The team investigate a machine learning problem they call ‘estimating the maximum’ (EMX), in which a website seeks to display targeted advertising to the visitors that browse the site most frequently – although it isn’t known in advance which visitors will visit the site. This problem is similar to a mathematical paradox called the continuum hypothesis, another field of investigation for Gödel.

Like the incompleteness theorems, the continuum hypothesis is concerned with mathematics that cannot ever be proved to be true or untrue, and given the conditions of the EMX example, at least, machine learning could hypothetically run into the same perpetual stalemate. (1)

We believe in science.

But the only thing science has proved is that it cannot prove anything.

Gödel’s incompleteness theorem showed that whatever we do, we will never be able to prove everything in the context of our limited theories. No matter how many axioms you choose and how carefully you choose them, you will never be able to describe the cosmos in its totality in an objective way as proponents of scientism would like to believe.

Now we have built big computers.

With the hope that they will answer everything.

But they cannot answer anything.

Because there is nothing to answer in the first place.

In a limited world there is no reason to analyze the Monad.

In an immeasurable cosmos you cannot count beyond One.

The maximum is zero.

The minimum is infinite.

A computer struggling to make sense of the problem. A man standing beside the computer trying to make sense of the computer. A bird flying by. Poor man…

Destroyers of the Heavens… (Musk vs. Astronomy)

Some of you might have heard that the skies will never be the same again…
If not, I would suggest you follow up the Space X micro-satellites story more closely.

Elon Musks’s company has decided to launch a constellation of 42,000 micro-satellites in the skies, so as to provide Internet services to everyone on the planet. This constellation will be called StarLink and has caused astronomer’s to have nightmares for many days now. So far Musk’s company has launched 120 of those satellites (the actual number doen’s play a role, since it will change as we speak). To understand the magnitude of the endeavor, note that from the beginning of the space age until now, we have put around 8,000 satellites in orbit! But the problem is not just a problem of potential space junk and troubles in handling and managing the satellites traffic in space (ESA has already once diverted one of its satellites to avoid collision with a StarLink satellite). By the way, note that also other companies like Amazon are planning to put into the skies similar constellations of satellites.

The main problem arises from the fact that the satellites are too close to Earth (around 550 km) and too reflective (and thus, too bright). This causes extreme problems for astronomy. Telescopes are used to avoid satellites in orbit when taking photographs of the sky. However this would be one satellite at some point interfering with one photo. Now there are dozens (and in the future thousands) of such satellites literally destroying any observations astronomers try to make, by leaving a bright trail of light in the photographs astronomers take. The picture below is indicative of the problem.

StarLink satellites leaving trails of light in a telescope photo

Similar problems arise when amateur astronomers take long-exposure photos of the sky…

And last but not least, at their peak, the StarLink constellation will be visible with the naked eye and whenever you look at the stars it will be like you are looking at a video game with many small dots passing fast through the dark sky between the stars. (actually the total number of StarLink satellites is much higher than the number of visible stars which is in rough approximation 9,000)

Note: Musk is trying to find a solution to the problem, but of course after it is created. He already launched a satellite which was painted black on its belly, so as to evaluate whether this could be a solution to the problem.

So there you go my dear friends.

We ruined the Earth. And now we will ruin the sky as well.

Take a look at the sky to-night. For it will never be the same again. And then go back to Earth. And plant a tree. Cover your hands in dirt. This is your way up to the stars. Stars which you saw perfectly long before there were any telescopes.

Heidegger had said so a long time ago.

The problem with technology is not that it doesn’t work.

But that it does…

Related articles & sources

Ice loss. From above…

Decades of satellite monitoring reveal Antarctic ice loss. (1)

People sitting in the warmth of their home.

Never to step on ice at all.

Knowing all there is to know about ice…

Detecting ice changes from above…

Heidegger once said it pretty eloquently.

The problem with technology is not that it doesn’t work.

But that it does…

Self driving cars. Selfless humanity. [Against Perfection]

Self-driving cars are taking to the streets in California this summer, but the Golden State isn’t the only one opening its roads to autonomous cars.

Virginia just announced that 70 miles of highway in the Commonwealth would be open to self-driving cars, like the cars in Google’s fleet. Any autonomous vehicle wanting to travel those routes, called the Virginia Automated Corridors, will be overseen by the Virginia Tech Transportation Institute, which helped the state government plan the project.

As the Richmond Times-Dispatchreports the plan is for companies to test how their cars react in real-world situations on highways packed with human drivers. People worried about potential crashes should remember that in California, accidents involving self-driving cars have all been the fault of humans so far. (1)

Yes, computers can drive better.

Computers can play chess better.

Cars can run faster.

So?

Lost our soul have we, to believe these things are important, as Yoda would say.

Start walking again. Start losing. Start playing chess for fun. Start making mistakes again. And you will soon feel that these things are not important. Being entails losing, imperfection. This is what makes us humans. This is what makes us perfect.

Technology does NOT work! No really. Ask your loan mower…

Many people argue in favor of our way of life. And the greatest and easiest way of arguing is throwing away lines like “Technology works”. Well, I was recently watching a video were a man with a scythe competed against a loan mower.

And guess what: the man with the scythe won! Not only did it complete the job faster, but it also worked as an “exercise” for the man who used it. Two in one! Way to go scythe!

Many times we see technology as something which is good but which has some drawbacks, i.e. alienating people, dehumanizing society, etc. But the fact is that technology does not actually work like we believe it does!

As in the example of the mower vs. scythe above, how many other examples are there out there where we have been FOOLED INTO believing that modern technology is better than the old?

Again I was watching in Youtube a video where an old tractor of 18 hp competed against a brand new one with a horsepower of 850. Guess who won the contest…

Yes, you guessed right! The OLD tractor won! What?!? It seems torque is something important when working on a field and not just being shinny, new and… fast. 😉

Examples keep coming up.

We all remember how that pigeon won the Internet in relaying information some years ago. In South Africa a Durban IT company pitted an 11-month-old bird armed with a 4GB memory stick against the ADSL service from the country’s biggest web firm, Telkom. Winston the pigeon took two hours to carry the data 60 miles – in the same time the ADSL had sent 4% of the data. (source)

Another interesting example is this: A soldier in full gear bear the same load as a medieval armored fighter. Which one will perform best on the run? Well, if you bid on the modern soldier you would lose. The weight of a late medieval full armor is around 30 kg (similar to that of the weight of the backpack of the soldier), but the burden is well distributed over the body. Thus, the medieval soldier is more flexible! (source)

Sure I know what you are going to say.

That these are specific examples, that in GENERAL technology works, that the Internet is faster in USA or in Europe than in Africa and so on and so forth. But have you put your BELIEF to the TEST? Would you expect that you should sell your mower and do everything by hand if you hadn’t seen that video? We should know better than to talk to unknown people via the Internet and call this “communication”. More… “primitive” people like in ancient Greece talked face-to-face every day in the Agora. And that worked pretty well for them…

Try to walk a steep hill on a street paved with asphalt. Now try to do the same on a steep paved with… nothing. By just walking on the uneven ground. Yes, I know. Walking on the “primitive” soil is easier and feels better.

Try to climp a steep bumpy hill with a car. Try to do the same with a horse. Yes I know. You do not want to ruin your car. But the horse is already up there on the top watching at you.

What else could you do better without technology?

As Arkas said, in the old days we had to get up from the couch to change the channel on the TV. Now we have to get up and walk the room 5 times to find the TV control.

Levi Strauss pointed out some time ago the fallacy of our times: What is “new” is not by default better than what is “old”.

Test the obvious and it will soon become awkwardly wrong…

Exit mobile version
%%footer%%