news

한어Русский языкEnglishFrançaisIndonesianSanskrit日本語DeutschPortuguêsΕλληνικάespañolItalianoSuomalainenLatina

New Intelligence Report

Editor: Editorial Department

【New Wisdom Introduction】AlphaGo's victory "is a watershed moment in the history of artificial intelligence" and also a "watershed" in Lee Sedol's life. That historic failure brought Lee Sedol's entire world down, but he is rebuilding his world in another way, while giving more people a warning - how to face the future when artificial intelligence is "ubiquitous."

Eight years after losing to AlphaGo, the former Go legend still hasn’t gotten over that dark moment…

In 2016, Korean chess player Lee Sedol lost 1:4 to DeepMind's strongest Go AI - AlphaGo

After the "God-killing" battle of artificial intelligence, Lee Sedol retired from the world, but the three words "Lee Sedol" became an unavoidable name in the history of the development of artificial intelligence, and are closely tied to AlphaGo.

A game that made the whole world talk about it for "a lifetime", and AI cast a long shadow on Lee Sedol's life.

If we say, "Anything that cannot defeat me will only make me stronger", then we can also say that only those who have truly experienced failure will understand the strength of their opponents. Lee Sedol is such a person.

Now, 41-year-old Lee Sedol has appeared in front of the media again and was interviewed by the New York Times. How will he rebuild the world after being defeated by AlphaGo, and how will he view artificial intelligence?

My whole world collapsed.

Lee Sedol, an 18-time world champion, is revered for his intuitive and creative playing style when playing the game of Go.

The Go genius's father is also a Go enthusiast and a teacher.

Influenced by his family, Lee Sedol started playing Go at the age of 5, and his talent was obvious from the beginning.

He soon became a leader among chess players of his age, not only in his local area but also in the whole of China, Japan and South Korea, and became a professional chess player at the age of 12.

At the age of 20, Lee Sedol had reached the highest level of Go, 9th dan, becoming one of the best players in the world and was described by some as the Roger Federer of Go.

"He's an idol, a star," said Lee Hajin, a former professional Go player. "Everyone looks up to him."

As Lee Sedol's status continued to rise, Go began to attract the interest of a new audience: computer scientists.

The temptation that Go brings to artificial intelligence researchers is huge, and they also want to participate in the world's most complex board game.

Go is much more complex than chess. It is often said that there are more possible variables on a Go board than there are atoms in the universe - the complexity of the game tree reaches 10 to the 360th power.

Humans are the only known species to have the cognitive abilities required to play Go, an ability that nature has spent hundreds of thousands of years evolving.

Artificial intelligence, by contrast, has achieved this goal in less than 20 years. DeepMind built AlphaGo using so-called neural networks, mathematical systems that can learn skills by analyzing vast amounts of data.

First, DeepMind fed the network 30 million moves from top chess players. Then, the program played game after game with itself until it learned which moves were successful and developed new strategies.

At the end of 2015, AlphaGo defeated the three-time European Go champion five times in a row in a closed-door match.

Then, Ms. Li, a former professional Go player working for the International Go Federation, approached Lee Sedol and proposed an open match with a prize of $1 million for defeating AlphaGo.

Lee Sedol said he accepted the proposal without giving it much thought, thinking it would be "very interesting" and not considering the possibility that he might fail.

The best-of-five match in Seoul was truly spectacular, broadcast on TV every night and watched by more than 200 million people, sparking widespread discussion and attention in our country.

World Go champion Ke Jie was later defeated by AlphaGo

During the game, a DeepMind engineer sat across from Lee Sedol and placed the pieces according to the information conveyed to him by AlphaGo.

Lee Sedol said that it was disturbing not to have a real human opponent. AlphaGo's chess playing style was something he had never seen before, and he could not interpret the thoughts and feelings of a machine.

In the past, playing Go was a game between people, which was not only rational but also involved a lot of emotional exchanges.

Go is also called "hand talk". As black and white fight back and forth, the changes in the rhythm of placing pieces and the strength of placing pieces can reflect the mental state of the players, just like communicating with sign language in the game.

But when playing against a machine, there is no such thing as emotion. He himself said that it is equivalent to looking at the standard answer, which is not like playing against a human.

The world watched in awe as AlphaGo cornered him and pulled off moves that no human player could have imagined. AlphaGo won four of the five games. His brother, professional Go player Lee Sang Hoon, still remembers thinking only: “This is impossible.”

"I was very surprised because I never thought I would lose," Mr. Lee Sedol said at the post-match press conference. "I didn't know AlphaGo would play such a perfect game of Go."

The impact of Lee Sedol's defeat went far beyond the game itself, with the shocking loss making headlines around the world.

It seemed like a clear sign that artificial intelligence was entering a deeply unsettling new era — AlphaGo’s victory demonstrated the limitless potential of AI to master superhuman skills once thought too complex for machines.

Lee is now convinced that humans can no longer compete with computers at Go. He says artificial intelligence has changed the nature of the game that originated in China more than 2,500 years ago.

Lee Sedol recently admitted in an interview with The New York Times, "Losing to artificial intelligence, in a sense, means that my whole world has collapsed."

"I can't get used to it. I think artificial intelligence will beat humans one day, I just don't think it's here yet."

Reconstruction after collapse

Lee Sedol sees Go as an art form that is closely tied to his character and an extension of his personality, but has been crushed by the efficiency of algorithms.

Unable to accept defeat, he resigned in 2019, believing that artificial intelligence was invincible and that he would always be second.

Lee Sedol is in the biggest crisis of his career. He is a more formidable opponent than Ke Jie. His emotionless moves are ruthlessly destroying human confidence.

However, there is no regret in making the move. Although the haze has shrouded the game for 8 years, when Lee Sedol looks back and talks about the man-machine battle, he feels more confident and guided to dispel the haze.

At a time when artificial intelligence is having an increasingly profound impact on the future of humanity, Lee Sedol warns us not to be caught off guard like him, but to understand the technology, become familiar with it, and apply it.

He has delivered many speeches on artificial intelligence, using his own personal experience to warn people not to underestimate technological progress and not to fight unprepared battles in more future human-machine games.

Since his failure, Lee Sedol has become an enthusiast of artificial intelligence, closely following the breakthroughs it brings.

AI is helping chatbots hold almost human-like conversations; it’s solving problems that have baffled scientists for decades, like predicting the shape of proteins; and it’s blurring the lines of creativity: independently composing music, making artwork, and generating videos.

Lee Sedol is not a doomsday theorist for artificial intelligence. In his view, artificial intelligence may replace some jobs, but it may also create some jobs.

When talking about AI and Go, he said, “It is important to remember that it was humans who both created Go and designed the AI ​​systems that mastered Go.”

On the other hand, Lee Sedol is also worried that artificial intelligence will change the core values ​​of human beings.

“In the past, people were in awe of creativity, ingenuity and innovation,” he said. “But since the advent of artificial intelligence, a lot of that has disappeared.”

When talking about the reason for retirement, Lee Sedol said sincerely, "I can no longer enjoy the game, so I retired."

In fact, Lee Sedol has been actively involved in the development of Go in another way.

He has written several books, including an autobiography and a series of books about his game.

He created his own board games inspired by Go, including Kingdom, Crown, and Nine Riders.

He also founded a Go academy for children, with more than a dozen branches across the country.

His 17-year-old daughter is in her last year of high school, and when they discuss what she should study in college, artificial intelligence often comes up.

He advised his daughter to study fields that are difficult to replace with artificial intelligence, and to choose a career that suits her in the tide of the times and the wave of technology.

“We often talk about choosing a job that won’t be easily replaced by AI or will be less affected by AI,” he said. “It’s only a matter of time before AI becomes ubiquitous.”

Lee Sedol's younger brother also said, "Artificial intelligence is affecting the world in an alarming way." He is still a professional chess player, and like other professional chess players, he now uses an artificial intelligence system for training.

“Professional players are figuring out how these algorithms work and trying to close the gap,” his brother said, “but we’re still a long way off.”

The man vs. machine war is a thing of the past

Fellow chess players Garry Kasparov and Magnus Carlsen were also unable to compete with AI, but they had a more positive view of it.

They say the board-playing AI AlphaZero, the successor to AlphaGo Zero, has advanced the game of chess by offering many novel moves.

Carlsen even regards AlphaZero as his idol.

There are more and more AI chess systems like AlphaZero that are more powerful than AlphaGo, which defeated Lee Sedol.

Faced with this situation, Kasparov is not worried about losing his job. He confidently stated that artificial intelligence has not fundamentally destroyed chess, and even if machines are better, people are still interested in human chess players.

At present, the reasoning capabilities of systems such as AlphaZero have not yet reached the level of LLM and generative AI, and they are still special-purpose AI.

However, DeepMind co-founder Demis Hassabis hinted that future Gemini models will achieve this goal, and OpenAI is also working to improve the logical capabilities of its models.

As AI becomes more and more intelligent, the story of "man vs. machine" eight years ago seems to no longer apply to today. If we regard artificial intelligence as an opponent, we may be defeated. But if artificial intelligence no longer exists as the opposite of humans, players who put aside the struggle for power may enjoy the purity of Go more.

In the future when artificial intelligence is "ubiquitous" as Lee Sedol has said, we will eventually move from "man vs. machine" to human-machine symbiosis.

References:

https://the-decoder.com/former-go-champion-lee-sedol-still-seems-to-be-struggling-with-ai-defeat/

https://www.nytimes.com/2024/07/10/world/asia/lee-saedol-go-ai.html