Kunal Sen was born in a creative household in Calcutta, India, raised by his film director father and actress mother. After completing his graduate studies in Physics, Kunal developed an interest in Computer Science and returned to graduate school. He moved to Chicago to obtain a Ph.D. in Artificial Intelligence. After earning his degree, he immersed himself in digital technology, designing electronic medical systems and writing technical books. He has led the technology group at Encyclopædia Britannica since 1999. In 2011, during a trip to India, Kunal decided to invest more in his interest in visual art, committing a significant amount of his time to artmaking. He had been thinking deeply about the convergence of science, philosophy, and art, seeking to express his understanding through his creations. His practice goes beyond traditional art forms, combining his expertise in electronic technology, programming, and painting to convey intricate narratives more effectively. This process has rendered his work increasingly personal, exploring universal aspects of humanity.
A fascinating account of the development of artificial intelligence is presented in this article “Thinking Machines”, which is framed by the technological and historical context of chess. The author persuasively contends that the most cutting-edge technologies of our day have continuously been contrasted with our comprehension and development of intelligent machines. The article starts off by stating that the human brain is the most intricate object in existence. It is a 1.5-kilogram soft tissue wonder that can think deeply, be creative, and be self-aware. From complex clockwork mechanisms and industrial engines to telephone networks and, eventually, the contemporary computer, Sen then explores the historical parallels for the mind. He suggests that the computer could be a lasting analogy since it is a framework for processing information, which is consistent with how we think the mind works. The historical quest to create computers that can mimic human intelligence occupies a large amount of the article, with chess being cited as a prime example of a task requiring high intelligence. The “Mechanical Turk,” a 1770 chess-playing automaton that enthralled audiences for 84 years and even vanquished famous people like Benjamin Franklin and Napoleon Bonaparte, is an intriguing historical narrative. Later on, it was discovered to be a complex deception that included a human player hidden inside its equipment. A compelling example of our enduring interest in sentient robots may be found in this story. The story then moves to the present day, when artificial intelligence started to take shape in earnest in the middle of the 20th century. The author emphasises how the concept of “intelligence” itself is changing, pointing out that once-intellectual skills like maths were pushed to the domain of algorithms as machines became more proficient in them. Alan Turing’s paper “Computing Machinery and Intelligence,” published in 1950, marked a turning point in this trip since it established the Turing Test as a standard for machine intelligence. The 1956 Dartmouth Conference was another significant occasion that established the notion that machines might replicate every facet of human intelligence. The article skilfully illustrates the development of AI problem-solving using the game of chess. Sen estimates that the game-tree complexity of chess is 10^123, which explains its enormous complexity. Both early AI programs and humans used three crucial shortcuts to address this:
1. Limited Search Depth: Instead of looking at the complete game tree, only a few move-countermove pairings are analysed.
2. Pruning: At each step, specifically examining the most promising moves.
3. Heuristic Evaluation: Applying general guidelines to determine whether board positions are favourable after a search.
The evolution of computers that can play chess provides a clear chronology of AI advancements. It took twenty years for a program to reach a Master-level, starting with IBM’s first complete program in 1957. The development of chess-playing computers serves as a clear timeline of AI progress. From the first full program by IBM in 1957, it took two decades for a program to reach a master-level rating and another decade for IBM’s Deep Thought to defeat a Grand-master. The watershed moment came in 1996 when IBM’s Deep Blue defeated world champion Garry Kasparov, signalling the end of human dominance in the game. The article points out that this victory was largely due to “brute force,” with Deep Blue capable of evaluating 200 million board positions per second. More recent engines like Stockfish, while still relying on massive searches, are more efficient at judging positions. Modern engines like Stockfish are better at determining locations, even if they still do extensive searches. Setting the foundation for a second portion, the article’s concluding parts promise to examine the transition from these algorithmic approaches to contemporary machine learning systems. As Sen suggests, instead of depending on human-programmed rules, these newer systems can learn to play sophisticated games like Go that were out of reach for older approaches by watching or competing with themselves. The generality of these learning systems—the same computer that learns to master chess may also learn other games through experience—is a crucial difference that is emphasised. In essence, “Thinking Machines” provides a well-structured and engaging overview of the historical and conceptual development of artificial intelligence. By using the universally understood and intellectually respected game of chess as a consistent thread, Kunal Sen successfully demystifies complex ideas and provides the reader with a clear appreciation for the field’s progress and future direction.