Sir Potata: Hi again; I’ll be taking over from here. Thanks to Kuyta, we now have a good idea of what AI is. But I think that most of us have already heard about it and know something about it (if you didn’t, then I’m afraid you are probably living in a cave). Recently, there has been a huge craze about Dall-E and ChatGPT. (God, don’t get me started on the ChatGPT; I swear, if someone mentions it one more time, I won’t be gentle.)
Kutay: ChatGPT
Sir Potata:
Aaaaaaanyways, In the past two years or so, AI has become more widely known and easier to use. However, as Kuyta mentioned, it took a lot of time to get to where we are now. Let’s take a look at how AI came to be what it is today and follow its steps throughout history.
I wouldn’t have guessed, but the foundations of AI were laid by people who preceded the technology by at least 200 years. But even before that, people who lived during the ancient times already dreamed of the concept in some way. For example, the bronze giant automaton Talos, the guardian of Crete Island, is based on a Greek myth. While this example is on the imaginative side of things, the logic behind machine learning is also being established. Around 350 BCE, Aristotle introduced the idea of syllogistic logic, a deductive logic system, in his book Prior Analytics. During the 17th century, three mathematicians—Leibniz, Hobbes, and Descartes—worked on expressing thoughts systematically.
Enough with the ancient history; let’s fast forward to times where computers exist, where it all began for real. We are in 1950, Alan Turing published Computing Machinery and Intelligence. This is also where the previously mentioned Turing test is introduced. He discussed the idea of a machine thinking like a human in detail in this paper, including objections and possibilities. He concluded his paper with the following paragraph:
“We may hope that machines will eventually compete with men in all purely intellectual fields. But which are the best ones to start with? Even this is a difficult decision. Many people think that a very abstract activity, like playing chess, would be best. It can also be maintained that it is best to provide the machine with the best sense organs that money can buy and then teach it to understand and speak English.”
The amount of foresight this man possessed was truly astounding. Just two years after he made that statement, Arthur Samuel developed a program that was able to defeat humans in a game of checkers. He was also going to be the one to use the term “machine learning” in 1959. Or the infamous Deep Blue, the computer built by IBM that defeated the world chess champion Kasparov in 1976.
As for understanding the language, that part was achieved to some extent in the 1960s. In 1965, Eliza, a natural language processing program that could interact with the user using language, was created by Joseph Weizenbaum. The concept was rather interesting; Eliza is a psychotherapist. The program used keywords in the answers to come up with an answer. However, the amount of vocabulary was too limited, and the answers were rather cliche. Here is the conversation I had with Eliza for the lols:
Sir Potata: Again, for further explanation of Eliza you can watch this video.
Then there is SHRDLU, another natural language understanding program that could understand, reply, and perform simple tasks using simple sentences. It was developed by Terry Winograd in 1970. (You can see it in action here)
Back in 1956, the historic year for AI, two historic events happened in the field. The “field” previously had no name. John McCarthy coined the term “artificial intelligence” during the Dartmouth Workshop, where the pioneers of AI met up and charted the course for the next decade. Their objective was as follows: “An attempt will be made to find how to make machines use language, form abstractions and concepts, solve kinds of problems now reserved for humans, and improve themselves.” (Here is the full text of the proposal)
The second development was the Logic Theorist by Allen Newell, Herbert A. Simon, and Cliff Shaw. It is officially the first artificial intelligence program in history. The program was used to prove several theorems in the book “Principia Mathematica”. This led to the creation of the Information Programming Language (IPL), which will be important in the creation of AI programming languages.
Kuyta: What? AI programming languages? I’ve actually never heard of that.
Sir Potata: Just like there there are there are specific programming languages specialized for specific tasks, eg. HTML for web programming or python for data analysis, researchers eventually had to come up with an easier way to code AI. As a result of that McCharty came up with the AI programming language Lisp, using IPL as a foundation. Even though it was 1960 when this happened, after more than 60 years, it is still used in the field which if you ask me, is pretty impressive.
What is more impressive is how long this post is becoming. I was a fool for thinking I would be able to fit everything in a post. Obviously it was a mistake. It is fascinating to see something that recently has exploded into popularity to have such history. Next post I will pick up from where I’ve left and complete this topic, until then I’m out.
