Stritar's chronolog

Do you think we are slowly reaching the end (of science)?

Sat, 07 Jul 2012 09:59:15 GMT

This was great week for science. The scientists from CERN Large Hadron Collider finally proved with great probability that the Higgs boson particle exists. Not that any of us mortals truly understand what it means for the future of mankind, but it's supposed to be quite significant, so I won't argue with that. Science has come a long way, and while we take into account a few other interesting and revolutionary fields, such as Artificial Intelligence, Biotechnolooy, Nuclear Fission, Stem Cells, Genetics, etc., we must also consider the timeframe in which these discoveries did or will take place, in relation to the history of our planet and humanity.

Imagine this: if history of Earth would be represented by one standard year, first cells would appear by the end of February, life would move to dry land around December 1st, dinosaurs would appear on December 13th, modern mammals on December 27th. On the evening on December 31st, first hominids would evolve in Africa, and 10 minutes before midnight, Neanderthals would spread throughout Europe. Around minute to midnight, agriculture would be invented, the Roman Empire would collapse 10 seconds before new year's, and the last 2 seconds would be marked by the industrial era. (source)

Makes you think about our insignificance in the history of Earth (not even the whole universe). But we've come a long way in these few minutes. The steam engine was invented 250 years ago, and look at us now. All in two seconds of Earth's history!

I think it was Isaac Asimov who said modern computerization represents the fourth big revolution in communication, after speech, writing and printing. Speech was invented tens of thousands of years ago, writing thousands of years ago, printing hundreds of years ago, and information technologies decades ago. Yes, it goes exponentially faster with each iteration, because every new generation of communication allows information to travel faster and reach more people, besides the fact that each time, global population is a few times greater than before. This enables science and knowledge to evolve even on a faster pace. Thousands of universities and scientist are already taking into account what CERN had discovered and announced, developing theories even further, making experiments of their own.

If all goes well, I think there is only one possible scenario: we will soon come to the end. Find that particle, understand that impulse, define that force. And by soon I don't mean tomorrow, or even in our generation, but in no time compared to History. Hundred, even thousand years? Why not. Ten, hundred thousand years? Sure, we have time, what are a few "minutes" more. Imagine everything we'll discover if we don't fuck it up on a major scale and our society evolves in a similar fashion as it did before.

But what is waiting for us at the end? Well, I think we are all pretty much aware of that. "Why are we here?", "What do we have to do?", "Is this real life?" and that sort of shit that brought us here in the first place. I think science has a good chance of clearing these things up one day, and then we will find the ultimate answer (or question). And Science will become one with Religion and the whole situation will be quite ironic, concerning the traditionally opposed stance they have. The only question that remains is, will we live happily ever after then, or will we just find another, greater and even more complex system, where we will have to help our Gods find their Gods?

Did Google just admit Apple's Siri is the future of search?

Sun, 04 Dec 2011 15:21:19 GMT

I don't know if you saw The evolution of Google search video, which they've published a few days ago. You should, it's a cool movie, portraying the history of search and Google's vision of its future. But something went wrong. One of the punchlines of the video was a story from one of the engineers, who said that next-generation search engines will be able to answer complex questions such as the following:

"Hey, what is the best time for me to sow seeds in India given that monsoon was early this year?’"

A very legitimate question.

I don't know if you've tried out iPhone's new personal assistant, Siri. It's awesome in every bit. Not only does it have a state-of-the-art voice recognition, it's also packed with super smart artificial intelligence that supposedly allows you to ask crazy things things such as:

"Can you remind me to call my wife when I leave the office?"

Another very legitimate question.

And there's a strong resemblance there. Both requests are really abstract and probably require quite a bit of computational power to be understood by a program. They have nothing to do with mathematical or social ranking currently used by Google (search), they are all about Artificial Intelligence and semantic interpretation. And while Google currently doesn't provide (or at least market) services that would be able to understand such sentences, Apple does.

I've noticed quite a few articles saying concepts such as Siri are the future of search. It's obvious artificial intelligence will play a big role in this segment. Apple's already in. Even if their technology is not superior to Google's, who is also working on embedding AI into search, it's fully available today, and everybody knows it. Google should really be careful with such statements concerning their core business, Web search. Specially if they are competing against the marketing wizards of Apple, who know how to sell things even if they don't fully work.

Promoting a technology you don't have and your competition does? Stupid consumers such as myself might do something stupid.

UPDATE (5.12.2011): You can join the discussion on HackerNews.

Can you believe Watson got the question about Slovenia wrong on Jeopardy?

Wed, 09 Mar 2011 07:33:12 GMT

Slovenia made it to the spotlight again, for the first time after the soccer world cup (when Slovenia was trending topic on Twitter and top search on Google). This time, it happened because IBM's supercomputer Watson competed against human champions in the famous TV show Jeopardy. IBM's computers are known to destroy people in various challenges, Deep Blue beat the world champion Garry Kasparov in a chess tournament in 1997. But chess is simple for computers to play, because it is pure logic and mathematics – the capability of a player is determined by the number of operations and actions it can calculate in advance. But a quiz is a totally different story, where the biggest challenge is semantics – understanding the meaning of words.

In 1950, Alan Turing, one of the greatest pioneers of computing introduced the Turing test, a methodology that could separate humans from computers using a set of questions, some of them formed in such a way computer wouldn't be able to understand and answer them. There are many questions which can't be answered with pure logic, the one I remember from high school goes something like this:

"Jack attended Sally's party, bring a doll. What was the present?"

The catch is in the connection between party – (birthday) – present – doll, which can't be noticed without abstract thinking humans are capable of. And today's computers still face the same problem - even though Watson dominated Jeopardy, it failed miserably on the following question about Slovenia:

"As of 2010, Croatia & Macedonia are candidates but this is the only former Yugoslav republic in the EU"

Watson's computing capabilities and knowledge banks are huge, but a question and an answer so obvious to humans presented a huge problem. Watson surely knows which countries are EU members, but it obviously didn't understand the question, thinking it was asked about which country would be next to start negotiating for EU membership, answering Serbia. The right answer was, of course, Slovenia.

The video is also fascinating from the cultural point of view – and extremely creepy. Those who have watched (or read) "2001: A Space Odyssey" may have experienced a slight shiver and carefully waited if Watson would say it: "Hello Dave". Others might have enjoyed this science fiction presentation, but besides Watson's obvious advantage in being the fastest to answer the question, it's clear that computers are still far away from being intelligent. And hopefully they will stay that way.