Everything Everywhere Daily: History, Science, Geography & More - All About Algorithms

Starting point is 00:00:00 Whether or not you know it, you use them every day. You are actually trained to use them as a child. And much of the world we live in today is directly or indirectly a result of their use. While they are ideally suited for computers, they were actually first developed thousands of years ago. Learn more about algorithms, what they are, how they work, and how they impact the world today on this episode of Everything Everywhere Daily. What if your perceptions about the past were wrong? ThruLine is a podcast that takes you back in time to uncover the, the parts of the story that may have gone unnoticed.

Starting point is 00:00:46 It effectively turned day into night. And how it shaped the world now. Time travel with us every week on the Thuline podcast from NPR. An algorithm is actually a very simple concept. An algorithm is nothing more than a set of rules. The dictionary definition of an algorithm says it is, quote, a finite set of unambiguous instructions that given some set of initial conditions can be performed in a described sequence to achieve certain goals that has a recognizable set of end conditions.

Starting point is 00:01:21 End quote. That sounds complicated, but if you've ever seen or used a decision-making flow chart, that's an algorithm. An even simpler example of an algorithm that everyone is probably used is a cooking recipe. You take certain ingredients, you combine them in a certain way, you heat them for a set time and temperature, and you get an end result. At first glance, an algorithm doesn't even seem like that big of a deal. Recipes and sets of instructions are hardly world shattering. However, algorithms are a very big deal. The first evidence we have of algorithms comes from the ancient world. Every ancient culture that had some form of mathematics, used algorithms to do basic mathematical operations. The Babylonians, Egyptians, Indians, Chinese, and

Starting point is 00:02:02 Greeks, all used algorithms. If you remember back to my episode on prime numbers, the sieve of Eratosthenes was an algorithm that was used to discover prime numbers. The sieve is pretty simple. Start with two, remove all the numbers which are multiples of two, which is all the even numbers. Then go to three and remove all the multiples of three. And just keep doing this, and what's left over are all the prime numbers. The word algorithm actually has a really interesting history. It comes from a 9th century Persian mathematician by the name of Muhammad Ibn Musa El Choirizami. El Choirizmi has come up in several previous episodes.

Starting point is 00:02:36 He was the person who developed Islamic accounting, and he was the person who was responsible for the spread of the Hindu-Arabic system of numbers that we use today. Al-Qua Rizami was so important to the development of early mathematics that I really should do an entire episode on him. Anyway, Al-Qa-Rizami's book on the subject of using Hindu and Arabic numbers was translated into Latin, and the Latinization of his name became Al-Garizami. At first, the entire Hindu-Arabic decimal number system was known in English as the algorithm. However, it eventually became associated with the system of just doing arithmetic tick by aligning digits of the same power of 10. This is probably the same system you used to learn how to do addition when you were younger.

Starting point is 00:03:18 The algorithm was an early algorithm. This eventually evolved into a term to generally describe any set of mathematical rules. By the 19th century, algorithm became algorithm in English. When you were growing up, you learned how to do addition, subtraction, multiplication, and division using a very rote method with a pencil and paper. You would line up numbers, carry the values, and if you followed all the steps correctly, you wound up with the right answer. No one probably told you at the time, but you were learning an algorithm. There were a handful of mathematical algorithms, which were used for things like calculating the digits of pi,

Starting point is 00:03:54 but there were a lot of algorithms prior to the development of digital computers. Enough to give it a name, but not enough to really make a science out of it. Computers, it turns out, are really good at executing algorithms. You have a number of set inputs, a final. number of steps and a definite output. And computers can do it really fast. Even the simplest things you can do on a computer will involve some sort of algorithm. Let's take one of the first algorithms you learn when you're learning computer programming, a sorting algorithm. Let's say you have a bunch of words that you want to put into alphabetical order. How do you do it? There are actually many

Starting point is 00:04:30 different sorting algorithms, but let's just assume a very simple one for the sake of illustration. Take the first two words and compare them to see if they are in alphabetical order. If they are in alphabetical order, do nothing. If they are not in alphabetical order, then swap them. Next, move to the second and third words and do the same thing. Keep doing this comparing words that are next to each other until you get to the end. Then start from the beginning and run through all of the words again. Keep doing this until you can go through the entire list of words without swapping anything.

Starting point is 00:05:00 The end result will be a list of words in alphabetical order. As I said, there are many different sorting algorithms, so how do you know which one to use? What you would normally want to do is to use whatever algorithm is the most efficient, or in other words, the algorithm which can achieve the end result in the fewest number of steps. These low-level algorithms are found everywhere in computing. Every computer program is, at some level, just a very complex algorithm. It's simply a set of instructions. It isn't just running computer programs.

Starting point is 00:05:30 When you make some sort of request on the internet for a web page, the data is sent from one router to another router. There are a mind-bogglingly large number of routes between you and the web page you want to download. Determining that route is done via an algorithm. Algorithms can be found much higher up the food chain as well. In fact, algorithms drive much of the information that you get every day. Let's take, for example, a search engine like Google. You type in a search term and Google will quickly come back with a list of websites that, in theory,

Starting point is 00:06:00 should answer your question. There might be an astounding number of webpages that could possibly be a result for that search term. How does Google determine what results to show and in what order? Google has an incredibly complicated algorithm that, according to their public statements, has over 200 different variables that they look at. Google was founded based on an academic paper written by its founders Larry Page and Sergey Brin. Their idea was that each page on the internet could be given a rank based primarily on the number of links pointing to that page. They thought of a link as a type of vote. Someone linked to a page, and the person making the link was in some way vouching for the page they wanted to link to.

Starting point is 00:06:39 The core idea that they had was in and of itself sound. However, they soon discovered a massive problem. If everybody knew what the inputs were for the algorithm, it was possible to optimize your website such that it would rank high in Google searches. For example, about 20 years ago, one of the major parts of the Google Outer, algorithm was looking at the number of times a keyword was used on a page. Let's say you were searching for used cars in Chicago. Well, all a page had to do was to rank high was to simply put used cars in Chicago over and over and over at the bottom of the page. Some sites would hide it

Starting point is 00:07:12 by masking the text using the same color as the background. The value of ranking high for certain keywords was enormous, and as such, optimizing webpages for search engines has become a multi-billion dollar industry. It has become a constant game of cat and mouse with Google constantly updating and tweaking their algorithm and companies trying to figure out what will make them rank higher. Algorithms have also become vitally important to the functioning of social media sites. Every social media site has something known as a feed. Everyone has a feed which is unique to them based on the people, groups, and brands that

Starting point is 00:07:45 you follow. In the first iteration of most social media sites, the feed was nothing more than a reverse chronology of the posts made by the people you followed. That's it. It was an algorithm, but it was a really simple algorithm. Over time, however, the algorithm became much more complex. The algorithm shifted from seeing everything which was published by the people you followed to a filtered feed based on an algorithm. The algorithm would determine what it thought you might be most interested in seeing based on how much interaction the post received and what you liked in the past. In theory, the way I described it, algorithm sounds rather innocuous, who wouldn't want to see content which was customized just for

Starting point is 00:08:25 them. However, the adoption of an algorithm-based system had serious unintended consequences. The algorithm served as a feedback loop. If you like something, you saw more of it, which increased the odds that you'd like it, reinforcing the algorithm sending you even more of the same type of content. It created a bubble, where the algorithm would only show you what it assumed you wanted to see and nothing else. I went to my name. niece's house recently, and she had the Netflix menu up on her television set. What she was shown was totally different from what I saw when I used Netflix, and I had no idea that most of the programs she was offered even existed. Some of the programs she was shown were

Starting point is 00:09:06 ones that I would watch, but that's hard to do when you don't even know they exist. It was all due to the Netflix algorithm, only showing me what it assumed I wanted to see. Similarly, there was a recent documentary on people who think that the Earth is flat, called Behind the Curve. In it, they attended a flat earth convention and asked people how they came to believe that the earth was flat. Almost every person said it started with a YouTube video. YouTube didn't have any nefarious plan to try to get people to think that the earth is flat. Nor did most of the people at the Flat Earth Convention go searching for information that the Earth was flat. They were led to these videos, bit by bit, from the YouTube algorithm.

Starting point is 00:09:45 Despite being in the 21st century, with all of the information in the world at our fingertips, There are more people today who think the earth is flat than there were 20 years ago, and this is mostly due to the unintended consequences of algorithms. When I was in high school, I spent an inordinate amount of time in the library of the local college. I would go and randomly read books, magazines, and academic journals. If you ever wondered how I wound up posting a podcast like this, in large part it started way back then when I was able to wander around a library reading things without everything being filtered through an algorithm.

Starting point is 00:10:16 In fact, one of the reasons behind the creation of this podcast was to give everyone who listens a chance to be exposed to topics they might otherwise never be exposed to because Internet algorithms would never show it to them. It's actually about the opposite of an algorithm, serendipity. Algorithms are a part of everyone's life nowadays. You can't really avoid it, and it isn't even necessarily a bad thing. The vast majority of them are very innocuous. However, algorithms also have a downside. They can wall you off from information by feeding you what they think you want to hear. The way around it is to just have a spot in your life for a little bit of randomness and serendipity.

Starting point is 00:11:00 Everything Everywhere Daily is an Airwave Media podcast. The executive producer is Darcy Adams. The associate producers are Thor Thompson and Peter Bennett. Today's review comes from listener James Viverka over at Apple Podcasts in the United States. He writes, More than bingeworthy. I absolutely love this podcast. It is so thoughtfully written, I have found myself listening for hours.

Starting point is 00:11:21 Thanks, James. Please let me know when you've joined the Completionist Club. I will get you the keys to the clubhouse and show you the secret handshake. Remember, if you leave a review over at Podchaser until the end of April, and there's only a few more days left, they will make a donation to help feed Ukrainian refugees, which will be matched by several other podcast companies.

Everything Everywhere Daily: History, Science, Geography & More - All About Algorithms

There aren't comments yet for this episode. Click on any sentence in the transcript to leave a comment.