Wednesday, November 22, 2017

Thankfully, Alphabet's chairman realizes Google can't determine which side is right

Plus some disillusionment from suggestions that the truth must be profitable

Willie Soon sent me a link to a yesterday's CNBC article,
Alphabet's Eric Schmidt: It can be 'very difficult' for Google’s search algorithm to understand truth
Eric Schmidt, a boss of Google's parent company, expresses his opinion that Google isn't capable of determining which side of news is right and which is fake when two sides vigorously and fundamentally disagree about something. Hopefully, it should also mean that Google won't try to play the "minister of truth" in a foreseeable future.

Ivo Lukačovič, the owner of Seznam.cz, a Czech local competitor of Google's, has been saying for years that he won't be dragged to politics etc. and be abused as a stamp of the truth by one side or another.

I think that Trump's victory has helped to calm down these worries about the Orwellian evolution in the U.S. 1984 may still be coming but I think it seems more distant now than it looked a year ago plus two weeks, before the latest U.S. presidential election. This optimistic description applies to the U.S. as a whole, not to some worst environments: for example, many universities already live in 1984.




Google's search engine is all about the ranking. Ranking of web pages in general. Ranking of web pages that contain some words, or a collection of sequences of words, or similar and equivalent phrases, and so on. Lots of quantities may be measured well by computer algorithms. One may estimate which pages talk about the same story, which pages respect the grammar, which pages use a colorful literary language, and so on.




Facebook is capable of figuring out which people have similar interests as you have, so you may be offered material that was found interesting by people who are just like you. It's easier to give you stories that were found interesting by your Facebook friends, and so on. Those things are easy. It's clear that these suggestions from "people like you" are generally preferred by the Facebook users and therefore the Facebook stockholders, too. Certain politicians and NGOs are dissatisfied with the nations as wholes and they would love to pump some politically correct material for everybody. Let's hope that freedom will defend itself at the end.

But then you may formulate questions – sometimes seemingly simple questions such as "Is Donald Trump a better president than Barack Obama?" or "Is string theory the most credible framework to go beyond quantum field theory?" or "Is the Bitcoin the money or the gold of the future?" or "Does quantum mechanics need observers?" and lots of others – and someone could demand Google to decide which pages are right and which are wrong or "fake news".



Well, it's just impossible. To answer the questions about quantum mechanics and string theory meaningfully, so that your answer isn't just another layman's guess etc., you really need to know a lot and be able to think a lot and you need lots of the human intelligence that can't be replaced with computer programs right now. If it could be replaced, we would do so.

In the video above, after 1:10, Feynman said he had an advantage to have learned how hard it is to really know something. How careful you have to be, and other conditions have to be satisfied. It's extremely difficult. If a Google-like algorithm could give us correct answers to those seemingly straightforward but difficult questions, scientists could indeed become unnecessary. They could be replaced with a collection of new Google servers.

It's clearly not possible. The political and other questions are probably less dependent on someone's deep thought or special expertise, and the answers are arguably more subjective, but in some broad sense, the situation is the same. Computers just can't replace the humans and their wisdom and intelligence. Equally importantly, it's crucial for the humans to keep their freedom to think different.

Concerning questions on string theory, over 99.9999% of the mankind just doesn't have a clue. This 99.9999% of the mankind, including a huge majority of authors of web pages and articles on the Internet, include lots of percent whose English and HTML layout etc. is better than that of the average string theorist. But it's obvious that all those random people's opinions are irrelevant in comparison with a string theorist's knowledge.

As I said, there are lots of things that a computer algorithm may determine rather well by analyzing a page – how much the author is connected to others, perhaps even how much the "system" is willing to pay to him or her, how well he or she has learned the English grammar, and so on. Some of these quantities can sometimes be used as proxies of the truly interesting questions about Trump or quantum mechanics or string theory etc.

However, none of these correlations can really be relied upon. If you're able to hire a writer whose prose is smooth, it doesn't mean that you're right. If you're rich, it doesn't mean that you're right. If you're connected, it doesn't mean you're right. If you have some control over the most influential or most visited servers, it doesn't mean that you're right. If you earn the support from lots of voters, it doesn't mean that you're right, either. In democracy, it gives you the power to govern but that's something else than being right.

I was shocked when someone basically told me that the truth you seriously believe is only something you can earn the money from. Holy cow, I was stunned. The truth has nothing to do with the money. The truth is often extremely modest, the truth is often a Cinderella – and lots of people and also rich people may often trample on the truth. Rich people sometimes bribe others in order to change what they overtly believe to be the truth. Some people may obviously be bought to "believe" what the payer wants them to believe. It's happening all the time, decent people consider these influences corrupt and unethical, and their unavoidable existence proves that the truth isn't the same thing as the money.

Why would someone link the truth to the money? It looks like a defining relationship behind complete corruption and immorality. One could identify the truth with the money only if he respects just the money – and the truth is just a word for him that may be helpful to earn the money. But if he earns the money in a different way, he's ready to restate his belief in the truth. Is that ethical? It's surely not scientific because the scientific truth doesn't change with the money.

I think you must understand my initial shock when someone who is wealthy and was apparently interested in science told me that he only respected the truth that can bring him a profit. If you can't make a profit out of something, you don't really believe it, I was told. If I am not ready to trade Bitcoin futures, I don't believe anything I write about the Bitcoin. Holy cow. The stupidity of joining this gambling is one important truth I am defending so why should I automatically join the trading? He refused a bet on whether the Bitcoin would be used to make lots of payments – and he clearly wanted to foist something on me, something that I didn't actually claim. I think the Bitcoin is more likely to collapse on December 11th than a huge fraction of the people think but I am not sure that the price will collapse on that day or week. Just because I can't make a profit out of something doesn't mean that it's not an important truth or that I don't strongly believe it, right? When it comes to the most important things, there's just no way to make profit out of them.

Why the hell someone would demand a profit to be a necessary condition for the truth? It's just so disgusting – and so self-evidently scientifically flawed. Such a person would probably label the sentence that "profit can't be made of the most important truths" to be self-contradictory because he probably defines the "important truth" as something that can make him profit. But if that's so, it just means that he has no respect for the truth for its own sake. He only cares about the money and uses the truth, or something pretending to be the truth, as a prostitute that can help him to earn the money.

When I thought about the disillusion differently, I saw another way to see it that is "less shockingly immoral" but still bad from my perspective. The idea that "only something that can bring you a profit is the truth" is an enhanced formulation of the Popperism, the falsifiability cult. Only falsifiable statements belong to science, these nut jobs constantly repeat.

The reality is that all propositions in science are falsifiable in principle but most of the deep ones almost certainly are extremely difficult or unlikely to be falsified or decided in practice, at least in a foreseeable future. That's how things are. The important truth simply can't be settled this easily. If some potentially deep question could be safely settled with a $1 million or even $1 billion experiment, it would probably not be too deep. Deep questions are just harder. You can't settle them easily. That is what makes them deep and what makes the questions important.

And you can't really make directly profit out of them. After all, no one makes any profit directly from pure science. The word "pure" really means that it is separated from applications and the "dirty" money that can be produced by applications. Does it mean that there are no truths and potential truths in pure science? I beg to differ. Almost no ordinary people care about pure science which means that they won't pay a penny to you but it doesn't mean that pure science is worthless.

We know some people who may even live on the "periphery of science" and who pay lip service to the Popperism and the demands that science has to be "practical". And the slogan that "the truth you really believe must make you rich" is just a slightly more disgusting cousin of this Popperism. This experience has strengthened my realization that the Popperists are extremely close to scientists waiting to be bought and corrupted. But yes, my relationship to someone who claims that "the truth you really believe must make you rich" probably had to deteriorate – and most likely irreversibly so because the conflict of this slogan with the very basic definition of integrity and morality as I understand them seems unbridgeable.

I sincerely hope that Google and Facebook won't try to enforce the "truth" on everyone as defined by some of the proxies which could easily mean that certain cliques could hijack the monopoly on truth on the Internet – they could "buy" the truth for something they have in abundance. But just like the trade with slaves, this "trade with the truth" should better be considered immoral or illegal. It shouldn't be possible to "buy the truth". Most people probably do believe and will believe lots of wrong things and those will spread among their friends and Facebook friends but I think it's clear that if someone gained the monopoly over the truth because he boasts the control over the money in an industry such as the media or something else, the results would be far worse. Freedom and democracy are the worst systems – except for all the others that have already been tried.

Those of us who value the truth and who respect its independence from the money are vastly more valuable as human beings than those who are ready to sell their truth to the highest bidder.

No comments:

Post a Comment