“Mechanical Turk” reference from OpenAI CEO: Continuing the illusion

Forum

“Mechanical Turk” reference from OpenAI CEO: Continuing the illusion

By admin, Aralık 13, 2024



See Full Size


OpenAI CEO’s Sam Altman, ChatGPTfrom account X after ‘s global crash for several hours. “I was about to offer to be the Mechanical Turk for ChatGPT.” he said. Of course, Altman is clearly making a joke here, but let’s state at the very beginning of the content: The subject has nothing to do with Turks or Türkiye. However, between the Mechanical Turk and artificial intelligence an ironic relationship is available.

Mechanical Turk: A chess illusion

Although throughout history people have tried to imitate their own intelligence and skills through machines, one of the first examples in this field emerged in the 18th century. It was designed by Austrian inventor Wolfgang von Kempelen in 1769 and “Mechanical Turk” The automaton known as (“The Turk” – “The Mechanical Turk”) was a concrete product of this dream and was actually built by the Empress Maria Theresa It was made for. The device, which remained intact until 1854 and participated in many tours, was destroyed as a result of a fire in the museum where it was last stored.

Ostensibly a machine capable of playing chess, the Mechanical Turk is both a technical genius and a it’s just an illusion It had a worldwide impact during its time with its structure. The Mechanical Turk consisted of a human-like figure working positioned at a table, wearing an Ottoman-style costume, and sitting at a chessboard.




“Mechanical Turk” reference from OpenAI CEO: Continuing the illusion



See Full Size


Kempelen claimed that the machine contained a mechanism consisting of complex gears and pulleys. The Mechanical Turk amazed the audience by often managing to defeat the opponents he faced. But the Mechanical Turk was an illusion. A human chess master hidden inside the machine controlled the figure and played it. The Mechanical Turk toured Europe and America throughout the 18th and 19th centuries, being exhibited to various royals, philosophers, and scientists. Napolyon Bonapart ve Benjamin Franklin He managed to defeat many historical figures in the game, including statesmen such as. in 1819 father of computer counted Charles Babbage He played two matches with and won both of them.

Babbage suspected that the automaton was not “intelligent” but merely an elaborate hoax, somehow concealing a person controlling Turk’s movements from the inside. Even though there was a human chess master inside the machine, those examining the machine could not access that compartment. It was seen that there were complex gear mechanisms in the opened parts. But still, Babbage was right. Meanwhile, Babbage, perhaps influenced by the Mechanical Turk, calculator (difference machine) and his invention made significant contributions to the development of today’s computers. Nowadays, computers are the only things that power everything, including artificial intelligence.

Although the Mechanical Turk was not a pioneer of the technological age, he represented the early ideas of artificial intelligence. These ideas emerged in the mid-20th century Alan Turing ve John McCarthy It would be transformed into a real scientific field with the work of scientists such as. And today, we live in the age of artificial intelligence.

Today’s illusion is “artificial intelligence”




“Mechanical Turk” reference from OpenAI CEO: Continuing the illusion



See Full Size


OpenAI Leading companies like GPT With its models, it is actually for us humansillusion” they continue to create. We use ChatGPT, today’s most popular artificial intelligence, as if there was a human on the other end and it was responding to what we wrote. The humanoid voice mode or the interfaces of the o1 model that “pretends to think” and the three-dot icons that reflect the perception of “writing” are just part of a successful illusion.

If we say illusion, we are of course not talking about a deception here. ChatGPT and similar “a true intelligenceHe is trying to create the perception. We constantly call these “artificial intelligence”. However, as we mentioned before, there is no “intelligence” yet; The illusion is already here. In this context, Sam Altman’s Mechanical Turk expression also has a basis.

We should note that Large Language Models (LLM) are behind GPT-4 or GPT-4o, which powers ChatGPT. All major language models operate on the basis of probabilities and statistical relationships rather than actual understanding. These are shaped by incredibly large data sets. All this data, wherever it comes from, is processed through a neural network consisting of multiple nodes and layers. So, the incredible power of LLMs is based entirely on the recognition and recreation of these patterns and possibilities, rather than any understanding of concepts or ideas

Although the sentences or images created by these artificial intelligences may be awe-inspiring, LLMs do not actually use words. They may not even know what words actually mean or what they might refer to in different situations; they see them as sets of interrelated numbers that present certain patterns and patterns. This indicates that the artificial intelligence we see and use is actually an illusion. Interestingly and ironically, we don’t actually know exactly how these language models work; just like those who didn’t know how the Mechanical Turk worked at that time. More interestingly, even those who make today’s artificial intelligence do not know how these models work.

For more details:






Source

https://en.wikipedia.org/wiki/Mechanical_Turk


https://spectrum.ieee.org/untold-history-of-ai-charles-babbage-and-the-turk


https://en.wikipedia.org/wiki/Charles_Babbage



















Etiketler: , , , , , ,
Araç çubuğuna atla