• The AI Pulse
  • Posts
  • šŸ§  AIā€™s Impact on the Chess World

šŸ§  AIā€™s Impact on the Chess World

PLUS: What Chessā€™s Run-in With AI Teaches Us About Human Creativity

Welcome back AI prodigies!

In todayā€™s ā€œšŸ§  Sunday Specialā€:

  • šŸ¤–AI Defeats Chess

  • šŸ‘‘Chess Defeats AI

  • šŸ–ŒReal Art is Thoroughly Human

  • šŸ”‘Key Takeaway

Read Time: 7 minutes

šŸŽ“Key Terms

  • Chess Engines: A computer program that analyzes chess positions to calculate the best chess move.

  • Generative AI (GenAI): Uses AI models trained on text, image, audio, video, or code to generate new content.

  • Large Language Models (LLMs): AI models pre-trained on vast amounts of data to generate human-like text.

šŸ©ŗ PULSE CHECK

Do you think AI-generated art will ever be viewed as true art?

Vote Below to View Live Results

Login or Subscribe to participate in polls.

šŸ¤–AI DEFEATS CHESS

By the turn of the millennium, the future of chess was under attack. In 1997, IBMā€™s ā€œDeep Blue Computerā€ accomplished something no machine had before. It defeated reigning world chess champion Garry Kasparov in a chess match under standard tournament conditions by evaluating 200 million chess positions per second. This defeat raised a provocative question: ā€œWould people still play, watch, and enjoy chess if a machine could beat the best player on Earth?ā€

šŸ‘‘CHESS DEFEATS AI

It turns out that people would still play, watch, and enjoy chess. Since 2001, online chess has grown exponentially. For example, in 2014, Chess.com had 10 million members. By 2022, Chess.com had over 100 million members. So, whyā€™s that? Itā€™s because IBMā€™s ā€œDeep Blue Computerā€ and todayā€™s most powerful Chess Engines like ā€œStockfishā€ lack something crucial: humanness. For all their processing power, they lack the human touch. Intuition is the spark of creativity that makes chess compelling. Computers are extraordinarily efficient but canā€™t replicate the unpredictable, innovative, and sometimes flawed decisions that make chess feel alive.

To address this lack of humanness, Google DeepMind developed ā€œAlphaZero,ā€ a computer program that taught itself from scratch how to master the game of chess by leveraging Reinforcement Learning (RL) to train its Neural Networks (NNs). ā€œAlphaZeroā€ played 44 million chess matches against itself within nine hours. This process was powered by three Tensor Processing Units (TPUs), which use Matrix Math (i.e., ā€œmxā€) to offer high-speed calculations. RL teaches a computer program to make decisions that result in the best outcomes. It mimics the ā€œtrial-and-errorā€ process humans use to learn, where actions lead to desired outcomes. NNs are designed to mimic the human brain.

ā€œAlphaZeroā€™sā€ self-teaching ability, RL, and NNs allowed it to make unconventional moves that resembled human-like tendencies. The New York Times (NYT) described ā€œAlphaZeroā€™sā€ play style as ā€œromantic and attacking.ā€ It didnā€™t just play better than humans; it played creatively. Despite this, the appeal of watching human versus human (i.e., ā€œhuman vs. humanā€) chess matches remained strong. Since 2018, chess viewership on Twitch has increased sixfold. In 2021, following Netflixā€™s release of The Queenā€™s Gambit, paired with the surge in online activity caused by the Coronavirus Disease (COVID-19), the most-watched tournament-style chess matches garnered 613,439 live viewers. This resilience of chess teaches us a lesson about art and human creativity in the age of GenAI.

šŸ–ŒREAL ART IS THOROUGHLY HUMAN

Chess isnā€™t considered art, but itā€™s a creative pursuit. Like art, it asks humans to make decisions while adhering to constraints. Although almost every pattern on the chess board has been produced before, every game produces novel combinations of openings, offense, and defense. Like chess players, all artists face constraints. Muralists cannot use more colors than they have on hand. Writers often abide by audience requests, bound to their cultural context. Filmmakers can only film with available cast members under a specified production budget. While AI applications operate under constraints like computing power, these obstacles are inherently disconnected from human experiences.

Magnus Carlsen, arguably the greatest human chess player of all time, knows he canā€™t beat ā€œStockfishā€ or ā€œAlphaZero.ā€ However, millions of chess enthusiasts still love watching him play. The appeal lies in not knowing heā€™s the absolute best, and watching him struggle, innovate, and persevere is a shared human experience. The paradox of AIā€™s potential to replace human creativity, yet its inability to do so, reveals the misdirection of our fears.

One possible explanation is that weā€™re not genuinely looking for perfection. AIā€™s relentless optimization makes it seem alien-like and detached from human experiences. When chess enthusiasts watch Kasparov or Carlsen, we see their mistakes, adaptations, and resilience. This unpredictability creates excitement and suspense. We donā€™t crave flawless performance; we crave the thrill of the unknown, anticipating a battle that could go either way. As Italian artist Leonardo da Vinci once remarked, ā€œArt is never finished, only abandoned.ā€ Creativity doesnā€™t peak at flawlessness; it thrives in the space between greatness and imperfection.

Yet thereā€™s a more straightforward interpretation of this paradox: we enjoy watching other humans. Weā€™d rather just watch Kasparov or Carlsen than an AI application simulating the perfect chess strategies. Although we may marvel at the perfect performance of a computer program, the limitations of being human in the moment resonate with us more. For instance, when we watch Carlsen play chess, we know heā€™s subject to our limitations. Like us, he makes mistakes, feels pressure, and pushes through. However, he achieves feats that stretch the boundaries of what we thought a human could accomplish.

This connection doesnā€™t exist when we watch ā€œStockfishā€ or ā€œAlphaZero.ā€ Thereā€™s an identification with Carlsen, a sense that we share a common human experience and that we, too, are capable of high achievements. We canā€™t relate to a computer programā€™s journey to success because it has no struggle, no path from novice to master, and no obstacles to overcome. This absence of a story makes computer programs less compelling to us. Just as we donā€™t compare Usain Boltā€™s speed to a cheetahā€™s speed or Michael Phelpsā€™s swimming to a dolphinā€™s swimming, we donā€™t view AIā€™s achievements on the same terms as a humanā€™s achievements.

šŸ’„BONUS TIP

In the short run, fears of AI applications replacing human creativity are unwarranted. In narrow, structured assessments, LLMs are quick, high-volume brainstormers. But an LLMā€™s ā€œreasoningā€ and ā€œcreativityā€ only reflect whatā€™s been digitized in the past and documented for the future. Who cares if LLMs outperform humans in chess strategies? Our preference for human achievement is rooted in the connection we feel with others who share our struggles, limitations, and vulnerabilities.

No matter how advanced an AI application is, it canā€™t replicate the depth of human experience. Well, for now. In the long run, future generations may only know a world in which AI applications thrive. Will they still prefer the human touch? Time will tell.

šŸ“’FINAL NOTE

FEEDBACK

How would you rate todayā€™s email?

It helps us improve the content for you!

Login or Subscribe to participate in polls.

ā¤ļøTAIP Review of The Week

ā€œSo great! Thatā€™s all.ā€

-Darius (1ļøāƒ£ šŸ‘Nailed it!)
REFER & EARN

šŸŽ‰Your Friends Learn, You Earn!

You currently have 0 referrals, only 1 away from receiving āš™ļøUltimate Prompt Engineering Guide.

Refer 3 friends to learn how to šŸ‘·ā€ā™€ļøBuild Custom Versions of OpenAIā€™s ChatGPT.

Reply

or to participate.